From 4925446fdb7dc5a2993b0c718a3576faaa60724f Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 13 Mar 2026 12:04:00 +0000
Subject: [PATCH 01/99] Refactor hybrid queries to use `alpha_param` and remove
 `0.7` default in favour of `None`

---
 weaviate/collections/grpc/shared.py           |  7 ++-
 .../queries/hybrid/generate/async_.pyi        | 26 ++++----
 .../queries/hybrid/generate/executor.py       | 28 ++++-----
 .../queries/hybrid/generate/sync.pyi          | 26 ++++----
 .../queries/hybrid/query/async_.pyi           | 26 ++++----
 .../queries/hybrid/query/executor.py          | 28 ++++-----
 .../collections/queries/hybrid/query/sync.pyi | 26 ++++----
 weaviate/proto/v1/regen.sh                    |  2 +-
 weaviate/proto/v1/v4216/v1/base_search_pb2.py | 62 ++++++++++---------
 .../proto/v1/v4216/v1/base_search_pb2.pyi     |  6 +-
 weaviate/proto/v1/v5261/v1/base_search_pb2.py | 62 ++++++++++---------
 .../proto/v1/v5261/v1/base_search_pb2.pyi     |  6 +-
 weaviate/proto/v1/v6300/v1/base_search_pb2.py | 62 ++++++++++---------
 .../proto/v1/v6300/v1/base_search_pb2.pyi     |  6 +-
 14 files changed, 195 insertions(+), 178 deletions(-)

diff --git a/weaviate/collections/grpc/shared.py b/weaviate/collections/grpc/shared.py
index 593811fb3..9a8949451 100644
--- a/weaviate/collections/grpc/shared.py
+++ b/weaviate/collections/grpc/shared.py
@@ -690,7 +690,12 @@ def _parse_hybrid(
             base_search_pb2.Hybrid(
                 properties=properties,
                 query=query,
-                alpha=float(alpha) if alpha is not None else None,
+                alpha=None
+                if self._weaviate_version.is_at_least(1, 36, 0)
+                else (alpha)
+                if alpha is not None
+                else None,
+                alpha_param=alpha if self._weaviate_version.is_at_least(1, 36, 0) else None,
                 fusion_type=(
                     cast(
                         base_search_pb2.Hybrid.FusionType,
diff --git a/weaviate/collections/queries/hybrid/generate/async_.pyi b/weaviate/collections/queries/hybrid/generate/async_.pyi
index a408a9696..182892fb9 100644
--- a/weaviate/collections/queries/hybrid/generate/async_.pyi
+++ b/weaviate/collections/queries/hybrid/generate/async_.pyi
@@ -42,7 +42,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -69,7 +69,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -96,7 +96,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -123,7 +123,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -150,7 +150,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -177,7 +177,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -204,7 +204,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -231,7 +231,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -258,7 +258,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -285,7 +285,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -312,7 +312,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -339,7 +339,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -366,7 +366,7 @@ class _HybridGenerateAsync(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
diff --git a/weaviate/collections/queries/hybrid/generate/executor.py b/weaviate/collections/queries/hybrid/generate/executor.py
index 9b6762a2b..408e434b1 100644
--- a/weaviate/collections/queries/hybrid/generate/executor.py
+++ b/weaviate/collections/queries/hybrid/generate/executor.py
@@ -53,7 +53,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -81,7 +81,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -109,7 +109,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -137,7 +137,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -165,7 +165,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -193,7 +193,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -223,7 +223,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -251,7 +251,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -279,7 +279,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -307,7 +307,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -335,7 +335,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -363,7 +363,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -392,7 +392,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -421,7 +421,7 @@ def hybrid(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
diff --git a/weaviate/collections/queries/hybrid/generate/sync.pyi b/weaviate/collections/queries/hybrid/generate/sync.pyi
index d2befe438..fc430f8f3 100644
--- a/weaviate/collections/queries/hybrid/generate/sync.pyi
+++ b/weaviate/collections/queries/hybrid/generate/sync.pyi
@@ -41,7 +41,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -68,7 +68,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -95,7 +95,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -122,7 +122,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -149,7 +149,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -176,7 +176,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -203,7 +203,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -230,7 +230,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -257,7 +257,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -284,7 +284,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -311,7 +311,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -338,7 +338,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -365,7 +365,7 @@ class _HybridGenerate(
         grouped_task: Union[str, _GroupedTask, None] = None,
         grouped_properties: Optional[List[str]] = None,
         generative_provider: Optional[_GenerativeConfigRuntime] = None,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
diff --git a/weaviate/collections/queries/hybrid/query/async_.pyi b/weaviate/collections/queries/hybrid/query/async_.pyi
index 96628ea81..855da7fc5 100644
--- a/weaviate/collections/queries/hybrid/query/async_.pyi
+++ b/weaviate/collections/queries/hybrid/query/async_.pyi
@@ -34,7 +34,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -57,7 +57,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -80,7 +80,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -103,7 +103,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -126,7 +126,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -149,7 +149,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -172,7 +172,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -195,7 +195,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -218,7 +218,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -241,7 +241,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -264,7 +264,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -287,7 +287,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -310,7 +310,7 @@ class _HybridQueryAsync(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
diff --git a/weaviate/collections/queries/hybrid/query/executor.py b/weaviate/collections/queries/hybrid/query/executor.py
index 7ed027abb..4c618f961 100644
--- a/weaviate/collections/queries/hybrid/query/executor.py
+++ b/weaviate/collections/queries/hybrid/query/executor.py
@@ -45,7 +45,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -69,7 +69,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -93,7 +93,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -117,7 +117,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -141,7 +141,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -165,7 +165,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -191,7 +191,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -215,7 +215,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -239,7 +239,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -263,7 +263,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -287,7 +287,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -311,7 +311,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -336,7 +336,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -361,7 +361,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
diff --git a/weaviate/collections/queries/hybrid/query/sync.pyi b/weaviate/collections/queries/hybrid/query/sync.pyi
index fcf5d559b..5d40db939 100644
--- a/weaviate/collections/queries/hybrid/query/sync.pyi
+++ b/weaviate/collections/queries/hybrid/query/sync.pyi
@@ -34,7 +34,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -57,7 +57,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -80,7 +80,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -103,7 +103,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -126,7 +126,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -149,7 +149,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -172,7 +172,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -195,7 +195,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -218,7 +218,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -241,7 +241,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -264,7 +264,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -287,7 +287,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
@@ -310,7 +310,7 @@ class _HybridQuery(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[HybridVectorType] = None,
         query_properties: Optional[List[str]] = None,
         fusion_type: Optional[HybridFusion] = None,
diff --git a/weaviate/proto/v1/regen.sh b/weaviate/proto/v1/regen.sh
index 4fcf2cb16..56b5391aa 100755
--- a/weaviate/proto/v1/regen.sh
+++ b/weaviate/proto/v1/regen.sh
@@ -9,7 +9,7 @@ SCRIPT_DIR="${0%/*}"
 cd "$SCRIPT_DIR"
 PROJECT_ROOT=$(pwd)
 # Get weaviate dir from arg or by navigating up from script location
-WEAVIATE_DIR="${1:-../../../../weaviate}"
+WEAVIATE_DIR="${1:-../../../../core/weaviate}"
 
 echo "Weaviate directory: $WEAVIATE_DIR"
 echo "Project root: $PROJECT_ROOT"
diff --git a/weaviate/proto/v1/v4216/v1/base_search_pb2.py b/weaviate/proto/v1/v4216/v1/base_search_pb2.py
index 5767fdb57..f263cb10c 100644
--- a/weaviate/proto/v1/v4216/v1/base_search_pb2.py
+++ b/weaviate/proto/v1/v4216/v1/base_search_pb2.py
@@ -14,7 +14,7 @@
 from weaviate.proto.v1.v4216.v1 import base_pb2 as v1_dot_base__pb2
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\xd0\x04\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\r\n\x05\x61lpha\x18\x04 \x01(\x02\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operator\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\xfe\x04\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\x11\n\x05\x61lpha\x18\x04 \x01(\x02\x42\x02\x18\x01\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x18\n\x0b\x61lpha_param\x18\x0c \x01(\x02H\x02\x88\x01\x01\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operatorB\x0e\n\x0c_alpha_param\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -26,6 +26,8 @@
   _VECTORFORTARGET.fields_by_name['vector_bytes']._serialized_options = b'\030\001'
   _HYBRID.fields_by_name['vector']._options = None
   _HYBRID.fields_by_name['vector']._serialized_options = b'\030\001'
+  _HYBRID.fields_by_name['alpha']._options = None
+  _HYBRID.fields_by_name['alpha']._serialized_options = b'\030\001'
   _HYBRID.fields_by_name['vector_bytes']._options = None
   _HYBRID.fields_by_name['vector_bytes']._serialized_options = b'\030\001'
   _HYBRID.fields_by_name['target_vectors']._options = None
@@ -56,8 +58,8 @@
   _NEARTHERMALSEARCH.fields_by_name['target_vectors']._serialized_options = b'\030\001'
   _NEARIMUSEARCH.fields_by_name['target_vectors']._options = None
   _NEARIMUSEARCH.fields_by_name['target_vectors']._serialized_options = b'\030\001'
-  _globals['_COMBINATIONMETHOD']._serialized_start=3337
-  _globals['_COMBINATIONMETHOD']._serialized_end=3575
+  _globals['_COMBINATIONMETHOD']._serialized_start=3383
+  _globals['_COMBINATIONMETHOD']._serialized_end=3621
   _globals['_WEIGHTSFORTARGET']._serialized_start=52
   _globals['_WEIGHTSFORTARGET']._serialized_end=102
   _globals['_TARGETS']._serialized_start=105
@@ -69,31 +71,31 @@
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_start=484
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_end=555
   _globals['_HYBRID']._serialized_start=586
-  _globals['_HYBRID']._serialized_end=1178
-  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1043
-  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1140
-  _globals['_NEARVECTOR']._serialized_start=1181
-  _globals['_NEARVECTOR']._serialized_end=1610
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1529
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1583
-  _globals['_NEAROBJECT']._serialized_start=1613
-  _globals['_NEAROBJECT']._serialized_end=1778
-  _globals['_NEARTEXTSEARCH']._serialized_start=1781
-  _globals['_NEARTEXTSEARCH']._serialized_end=2149
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2042
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2096
-  _globals['_NEARIMAGESEARCH']._serialized_start=2152
-  _globals['_NEARIMAGESEARCH']._serialized_end=2325
-  _globals['_NEARAUDIOSEARCH']._serialized_start=2328
-  _globals['_NEARAUDIOSEARCH']._serialized_end=2501
-  _globals['_NEARVIDEOSEARCH']._serialized_start=2504
-  _globals['_NEARVIDEOSEARCH']._serialized_end=2677
-  _globals['_NEARDEPTHSEARCH']._serialized_start=2680
-  _globals['_NEARDEPTHSEARCH']._serialized_end=2853
-  _globals['_NEARTHERMALSEARCH']._serialized_start=2856
-  _globals['_NEARTHERMALSEARCH']._serialized_end=3033
-  _globals['_NEARIMUSEARCH']._serialized_start=3036
-  _globals['_NEARIMUSEARCH']._serialized_end=3205
-  _globals['_BM25']._serialized_start=3207
-  _globals['_BM25']._serialized_end=3334
+  _globals['_HYBRID']._serialized_end=1224
+  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1073
+  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1170
+  _globals['_NEARVECTOR']._serialized_start=1227
+  _globals['_NEARVECTOR']._serialized_end=1656
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1575
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1629
+  _globals['_NEAROBJECT']._serialized_start=1659
+  _globals['_NEAROBJECT']._serialized_end=1824
+  _globals['_NEARTEXTSEARCH']._serialized_start=1827
+  _globals['_NEARTEXTSEARCH']._serialized_end=2195
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2088
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2142
+  _globals['_NEARIMAGESEARCH']._serialized_start=2198
+  _globals['_NEARIMAGESEARCH']._serialized_end=2371
+  _globals['_NEARAUDIOSEARCH']._serialized_start=2374
+  _globals['_NEARAUDIOSEARCH']._serialized_end=2547
+  _globals['_NEARVIDEOSEARCH']._serialized_start=2550
+  _globals['_NEARVIDEOSEARCH']._serialized_end=2723
+  _globals['_NEARDEPTHSEARCH']._serialized_start=2726
+  _globals['_NEARDEPTHSEARCH']._serialized_end=2899
+  _globals['_NEARTHERMALSEARCH']._serialized_start=2902
+  _globals['_NEARTHERMALSEARCH']._serialized_end=3079
+  _globals['_NEARIMUSEARCH']._serialized_start=3082
+  _globals['_NEARIMUSEARCH']._serialized_end=3251
+  _globals['_BM25']._serialized_start=3253
+  _globals['_BM25']._serialized_end=3380
 # @@protoc_insertion_point(module_scope)
diff --git a/weaviate/proto/v1/v4216/v1/base_search_pb2.pyi b/weaviate/proto/v1/v4216/v1/base_search_pb2.pyi
index bac1b47b8..ab8a08f38 100644
--- a/weaviate/proto/v1/v4216/v1/base_search_pb2.pyi
+++ b/weaviate/proto/v1/v4216/v1/base_search_pb2.pyi
@@ -67,7 +67,7 @@ class SearchOperatorOptions(_message.Message):
     def __init__(self, operator: _Optional[_Union[SearchOperatorOptions.Operator, str]] = ..., minimum_or_tokens_match: _Optional[int] = ...) -> None: ...
 
 class Hybrid(_message.Message):
-    __slots__ = ["query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "vector_distance", "vectors"]
+    __slots__ = ["query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "alpha_param", "vector_distance", "vectors"]
     class FusionType(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
         __slots__ = []
         FUSION_TYPE_UNSPECIFIED: _ClassVar[Hybrid.FusionType]
@@ -87,6 +87,7 @@ class Hybrid(_message.Message):
     NEAR_VECTOR_FIELD_NUMBER: _ClassVar[int]
     TARGETS_FIELD_NUMBER: _ClassVar[int]
     BM25_SEARCH_OPERATOR_FIELD_NUMBER: _ClassVar[int]
+    ALPHA_PARAM_FIELD_NUMBER: _ClassVar[int]
     VECTOR_DISTANCE_FIELD_NUMBER: _ClassVar[int]
     VECTORS_FIELD_NUMBER: _ClassVar[int]
     query: str
@@ -100,9 +101,10 @@ class Hybrid(_message.Message):
     near_vector: NearVector
     targets: Targets
     bm25_search_operator: SearchOperatorOptions
+    alpha_param: float
     vector_distance: float
     vectors: _containers.RepeatedCompositeFieldContainer[_base_pb2.Vectors]
-    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
+    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., alpha_param: _Optional[float] = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
 
 class NearVector(_message.Message):
     __slots__ = ["vector", "certainty", "distance", "vector_bytes", "target_vectors", "targets", "vector_per_target", "vector_for_targets", "vectors"]
diff --git a/weaviate/proto/v1/v5261/v1/base_search_pb2.py b/weaviate/proto/v1/v5261/v1/base_search_pb2.py
index cde241b1e..cae3e4745 100644
--- a/weaviate/proto/v1/v5261/v1/base_search_pb2.py
+++ b/weaviate/proto/v1/v5261/v1/base_search_pb2.py
@@ -15,7 +15,7 @@
 from weaviate.proto.v1.v5261.v1 import base_pb2 as v1_dot_base__pb2
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\xd0\x04\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\r\n\x05\x61lpha\x18\x04 \x01(\x02\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operator\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\xfe\x04\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\x11\n\x05\x61lpha\x18\x04 \x01(\x02\x42\x02\x18\x01\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x18\n\x0b\x61lpha_param\x18\x0c \x01(\x02H\x02\x88\x01\x01\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operatorB\x0e\n\x0c_alpha_param\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -27,6 +27,8 @@
   _globals['_VECTORFORTARGET'].fields_by_name['vector_bytes']._serialized_options = b'\030\001'
   _globals['_HYBRID'].fields_by_name['vector']._loaded_options = None
   _globals['_HYBRID'].fields_by_name['vector']._serialized_options = b'\030\001'
+  _globals['_HYBRID'].fields_by_name['alpha']._loaded_options = None
+  _globals['_HYBRID'].fields_by_name['alpha']._serialized_options = b'\030\001'
   _globals['_HYBRID'].fields_by_name['vector_bytes']._loaded_options = None
   _globals['_HYBRID'].fields_by_name['vector_bytes']._serialized_options = b'\030\001'
   _globals['_HYBRID'].fields_by_name['target_vectors']._loaded_options = None
@@ -57,8 +59,8 @@
   _globals['_NEARTHERMALSEARCH'].fields_by_name['target_vectors']._serialized_options = b'\030\001'
   _globals['_NEARIMUSEARCH'].fields_by_name['target_vectors']._loaded_options = None
   _globals['_NEARIMUSEARCH'].fields_by_name['target_vectors']._serialized_options = b'\030\001'
-  _globals['_COMBINATIONMETHOD']._serialized_start=3337
-  _globals['_COMBINATIONMETHOD']._serialized_end=3575
+  _globals['_COMBINATIONMETHOD']._serialized_start=3383
+  _globals['_COMBINATIONMETHOD']._serialized_end=3621
   _globals['_WEIGHTSFORTARGET']._serialized_start=52
   _globals['_WEIGHTSFORTARGET']._serialized_end=102
   _globals['_TARGETS']._serialized_start=105
@@ -70,31 +72,31 @@
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_start=484
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_end=555
   _globals['_HYBRID']._serialized_start=586
-  _globals['_HYBRID']._serialized_end=1178
-  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1043
-  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1140
-  _globals['_NEARVECTOR']._serialized_start=1181
-  _globals['_NEARVECTOR']._serialized_end=1610
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1529
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1583
-  _globals['_NEAROBJECT']._serialized_start=1613
-  _globals['_NEAROBJECT']._serialized_end=1778
-  _globals['_NEARTEXTSEARCH']._serialized_start=1781
-  _globals['_NEARTEXTSEARCH']._serialized_end=2149
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2042
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2096
-  _globals['_NEARIMAGESEARCH']._serialized_start=2152
-  _globals['_NEARIMAGESEARCH']._serialized_end=2325
-  _globals['_NEARAUDIOSEARCH']._serialized_start=2328
-  _globals['_NEARAUDIOSEARCH']._serialized_end=2501
-  _globals['_NEARVIDEOSEARCH']._serialized_start=2504
-  _globals['_NEARVIDEOSEARCH']._serialized_end=2677
-  _globals['_NEARDEPTHSEARCH']._serialized_start=2680
-  _globals['_NEARDEPTHSEARCH']._serialized_end=2853
-  _globals['_NEARTHERMALSEARCH']._serialized_start=2856
-  _globals['_NEARTHERMALSEARCH']._serialized_end=3033
-  _globals['_NEARIMUSEARCH']._serialized_start=3036
-  _globals['_NEARIMUSEARCH']._serialized_end=3205
-  _globals['_BM25']._serialized_start=3207
-  _globals['_BM25']._serialized_end=3334
+  _globals['_HYBRID']._serialized_end=1224
+  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1073
+  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1170
+  _globals['_NEARVECTOR']._serialized_start=1227
+  _globals['_NEARVECTOR']._serialized_end=1656
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1575
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1629
+  _globals['_NEAROBJECT']._serialized_start=1659
+  _globals['_NEAROBJECT']._serialized_end=1824
+  _globals['_NEARTEXTSEARCH']._serialized_start=1827
+  _globals['_NEARTEXTSEARCH']._serialized_end=2195
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2088
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2142
+  _globals['_NEARIMAGESEARCH']._serialized_start=2198
+  _globals['_NEARIMAGESEARCH']._serialized_end=2371
+  _globals['_NEARAUDIOSEARCH']._serialized_start=2374
+  _globals['_NEARAUDIOSEARCH']._serialized_end=2547
+  _globals['_NEARVIDEOSEARCH']._serialized_start=2550
+  _globals['_NEARVIDEOSEARCH']._serialized_end=2723
+  _globals['_NEARDEPTHSEARCH']._serialized_start=2726
+  _globals['_NEARDEPTHSEARCH']._serialized_end=2899
+  _globals['_NEARTHERMALSEARCH']._serialized_start=2902
+  _globals['_NEARTHERMALSEARCH']._serialized_end=3079
+  _globals['_NEARIMUSEARCH']._serialized_start=3082
+  _globals['_NEARIMUSEARCH']._serialized_end=3251
+  _globals['_BM25']._serialized_start=3253
+  _globals['_BM25']._serialized_end=3380
 # @@protoc_insertion_point(module_scope)
diff --git a/weaviate/proto/v1/v5261/v1/base_search_pb2.pyi b/weaviate/proto/v1/v5261/v1/base_search_pb2.pyi
index 5f1871ac7..ca6f234bc 100644
--- a/weaviate/proto/v1/v5261/v1/base_search_pb2.pyi
+++ b/weaviate/proto/v1/v5261/v1/base_search_pb2.pyi
@@ -67,7 +67,7 @@ class SearchOperatorOptions(_message.Message):
     def __init__(self, operator: _Optional[_Union[SearchOperatorOptions.Operator, str]] = ..., minimum_or_tokens_match: _Optional[int] = ...) -> None: ...
 
 class Hybrid(_message.Message):
-    __slots__ = ("query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "vector_distance", "vectors")
+    __slots__ = ("query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "alpha_param", "vector_distance", "vectors")
     class FusionType(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
         __slots__ = ()
         FUSION_TYPE_UNSPECIFIED: _ClassVar[Hybrid.FusionType]
@@ -87,6 +87,7 @@ class Hybrid(_message.Message):
     NEAR_VECTOR_FIELD_NUMBER: _ClassVar[int]
     TARGETS_FIELD_NUMBER: _ClassVar[int]
     BM25_SEARCH_OPERATOR_FIELD_NUMBER: _ClassVar[int]
+    ALPHA_PARAM_FIELD_NUMBER: _ClassVar[int]
     VECTOR_DISTANCE_FIELD_NUMBER: _ClassVar[int]
     VECTORS_FIELD_NUMBER: _ClassVar[int]
     query: str
@@ -100,9 +101,10 @@ class Hybrid(_message.Message):
     near_vector: NearVector
     targets: Targets
     bm25_search_operator: SearchOperatorOptions
+    alpha_param: float
     vector_distance: float
     vectors: _containers.RepeatedCompositeFieldContainer[_base_pb2.Vectors]
-    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
+    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., alpha_param: _Optional[float] = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
 
 class NearVector(_message.Message):
     __slots__ = ("vector", "certainty", "distance", "vector_bytes", "target_vectors", "targets", "vector_per_target", "vector_for_targets", "vectors")
diff --git a/weaviate/proto/v1/v6300/v1/base_search_pb2.py b/weaviate/proto/v1/v6300/v1/base_search_pb2.py
index cbf099302..bc7884f55 100644
--- a/weaviate/proto/v1/v6300/v1/base_search_pb2.py
+++ b/weaviate/proto/v1/v6300/v1/base_search_pb2.py
@@ -25,7 +25,7 @@
 from weaviate.proto.v1.v6300.v1 import base_pb2 as v1_dot_base__pb2
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\xd0\x04\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\r\n\x05\x61lpha\x18\x04 \x01(\x02\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operator\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\xfe\x04\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\x11\n\x05\x61lpha\x18\x04 \x01(\x02\x42\x02\x18\x01\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x18\n\x0b\x61lpha_param\x18\x0c \x01(\x02H\x02\x88\x01\x01\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operatorB\x0e\n\x0c_alpha_param\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -37,6 +37,8 @@
   _globals['_VECTORFORTARGET'].fields_by_name['vector_bytes']._serialized_options = b'\030\001'
   _globals['_HYBRID'].fields_by_name['vector']._loaded_options = None
   _globals['_HYBRID'].fields_by_name['vector']._serialized_options = b'\030\001'
+  _globals['_HYBRID'].fields_by_name['alpha']._loaded_options = None
+  _globals['_HYBRID'].fields_by_name['alpha']._serialized_options = b'\030\001'
   _globals['_HYBRID'].fields_by_name['vector_bytes']._loaded_options = None
   _globals['_HYBRID'].fields_by_name['vector_bytes']._serialized_options = b'\030\001'
   _globals['_HYBRID'].fields_by_name['target_vectors']._loaded_options = None
@@ -67,8 +69,8 @@
   _globals['_NEARTHERMALSEARCH'].fields_by_name['target_vectors']._serialized_options = b'\030\001'
   _globals['_NEARIMUSEARCH'].fields_by_name['target_vectors']._loaded_options = None
   _globals['_NEARIMUSEARCH'].fields_by_name['target_vectors']._serialized_options = b'\030\001'
-  _globals['_COMBINATIONMETHOD']._serialized_start=3337
-  _globals['_COMBINATIONMETHOD']._serialized_end=3575
+  _globals['_COMBINATIONMETHOD']._serialized_start=3383
+  _globals['_COMBINATIONMETHOD']._serialized_end=3621
   _globals['_WEIGHTSFORTARGET']._serialized_start=52
   _globals['_WEIGHTSFORTARGET']._serialized_end=102
   _globals['_TARGETS']._serialized_start=105
@@ -80,31 +82,31 @@
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_start=484
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_end=555
   _globals['_HYBRID']._serialized_start=586
-  _globals['_HYBRID']._serialized_end=1178
-  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1043
-  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1140
-  _globals['_NEARVECTOR']._serialized_start=1181
-  _globals['_NEARVECTOR']._serialized_end=1610
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1529
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1583
-  _globals['_NEAROBJECT']._serialized_start=1613
-  _globals['_NEAROBJECT']._serialized_end=1778
-  _globals['_NEARTEXTSEARCH']._serialized_start=1781
-  _globals['_NEARTEXTSEARCH']._serialized_end=2149
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2042
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2096
-  _globals['_NEARIMAGESEARCH']._serialized_start=2152
-  _globals['_NEARIMAGESEARCH']._serialized_end=2325
-  _globals['_NEARAUDIOSEARCH']._serialized_start=2328
-  _globals['_NEARAUDIOSEARCH']._serialized_end=2501
-  _globals['_NEARVIDEOSEARCH']._serialized_start=2504
-  _globals['_NEARVIDEOSEARCH']._serialized_end=2677
-  _globals['_NEARDEPTHSEARCH']._serialized_start=2680
-  _globals['_NEARDEPTHSEARCH']._serialized_end=2853
-  _globals['_NEARTHERMALSEARCH']._serialized_start=2856
-  _globals['_NEARTHERMALSEARCH']._serialized_end=3033
-  _globals['_NEARIMUSEARCH']._serialized_start=3036
-  _globals['_NEARIMUSEARCH']._serialized_end=3205
-  _globals['_BM25']._serialized_start=3207
-  _globals['_BM25']._serialized_end=3334
+  _globals['_HYBRID']._serialized_end=1224
+  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1073
+  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1170
+  _globals['_NEARVECTOR']._serialized_start=1227
+  _globals['_NEARVECTOR']._serialized_end=1656
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1575
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1629
+  _globals['_NEAROBJECT']._serialized_start=1659
+  _globals['_NEAROBJECT']._serialized_end=1824
+  _globals['_NEARTEXTSEARCH']._serialized_start=1827
+  _globals['_NEARTEXTSEARCH']._serialized_end=2195
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2088
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2142
+  _globals['_NEARIMAGESEARCH']._serialized_start=2198
+  _globals['_NEARIMAGESEARCH']._serialized_end=2371
+  _globals['_NEARAUDIOSEARCH']._serialized_start=2374
+  _globals['_NEARAUDIOSEARCH']._serialized_end=2547
+  _globals['_NEARVIDEOSEARCH']._serialized_start=2550
+  _globals['_NEARVIDEOSEARCH']._serialized_end=2723
+  _globals['_NEARDEPTHSEARCH']._serialized_start=2726
+  _globals['_NEARDEPTHSEARCH']._serialized_end=2899
+  _globals['_NEARTHERMALSEARCH']._serialized_start=2902
+  _globals['_NEARTHERMALSEARCH']._serialized_end=3079
+  _globals['_NEARIMUSEARCH']._serialized_start=3082
+  _globals['_NEARIMUSEARCH']._serialized_end=3251
+  _globals['_BM25']._serialized_start=3253
+  _globals['_BM25']._serialized_end=3380
 # @@protoc_insertion_point(module_scope)
diff --git a/weaviate/proto/v1/v6300/v1/base_search_pb2.pyi b/weaviate/proto/v1/v6300/v1/base_search_pb2.pyi
index 80abcb05d..9ad165a42 100644
--- a/weaviate/proto/v1/v6300/v1/base_search_pb2.pyi
+++ b/weaviate/proto/v1/v6300/v1/base_search_pb2.pyi
@@ -68,7 +68,7 @@ class SearchOperatorOptions(_message.Message):
     def __init__(self, operator: _Optional[_Union[SearchOperatorOptions.Operator, str]] = ..., minimum_or_tokens_match: _Optional[int] = ...) -> None: ...
 
 class Hybrid(_message.Message):
-    __slots__ = ("query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "vector_distance", "vectors")
+    __slots__ = ("query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "alpha_param", "vector_distance", "vectors")
     class FusionType(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
         __slots__ = ()
         FUSION_TYPE_UNSPECIFIED: _ClassVar[Hybrid.FusionType]
@@ -88,6 +88,7 @@ class Hybrid(_message.Message):
     NEAR_VECTOR_FIELD_NUMBER: _ClassVar[int]
     TARGETS_FIELD_NUMBER: _ClassVar[int]
     BM25_SEARCH_OPERATOR_FIELD_NUMBER: _ClassVar[int]
+    ALPHA_PARAM_FIELD_NUMBER: _ClassVar[int]
     VECTOR_DISTANCE_FIELD_NUMBER: _ClassVar[int]
     VECTORS_FIELD_NUMBER: _ClassVar[int]
     query: str
@@ -101,9 +102,10 @@ class Hybrid(_message.Message):
     near_vector: NearVector
     targets: Targets
     bm25_search_operator: SearchOperatorOptions
+    alpha_param: float
     vector_distance: float
     vectors: _containers.RepeatedCompositeFieldContainer[_base_pb2.Vectors]
-    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
+    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., alpha_param: _Optional[float] = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
 
 class NearVector(_message.Message):
     __slots__ = ("vector", "certainty", "distance", "vector_bytes", "target_vectors", "targets", "vector_per_target", "vector_for_targets", "vectors")

From 1899ffdb5959a50e69ae7fe1afcac87ab72654b8 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 12:12:49 +0000
Subject: [PATCH 02/99] Remove mistakenly commited local change to `regen.sh`

---
 weaviate/proto/v1/regen.sh | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/weaviate/proto/v1/regen.sh b/weaviate/proto/v1/regen.sh
index 56b5391aa..4fcf2cb16 100755
--- a/weaviate/proto/v1/regen.sh
+++ b/weaviate/proto/v1/regen.sh
@@ -9,7 +9,7 @@ SCRIPT_DIR="${0%/*}"
 cd "$SCRIPT_DIR"
 PROJECT_ROOT=$(pwd)
 # Get weaviate dir from arg or by navigating up from script location
-WEAVIATE_DIR="${1:-../../../../core/weaviate}"
+WEAVIATE_DIR="${1:-../../../../weaviate}"
 
 echo "Weaviate directory: $WEAVIATE_DIR"
 echo "Project root: $PROJECT_ROOT"

From c057270f76b495bafbc33ffe601452b4f482a2e0 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 12:13:51 +0000
Subject: [PATCH 03/99] Update logic to use new proto message

---
 weaviate/collections/grpc/shared.py           |  1 +
 weaviate/proto/v1/v4216/v1/base_search_pb2.py | 60 +++++++++----------
 .../proto/v1/v4216/v1/base_search_pb2.pyi     |  6 +-
 weaviate/proto/v1/v5261/v1/base_search_pb2.py | 60 +++++++++----------
 .../proto/v1/v5261/v1/base_search_pb2.pyi     |  6 +-
 weaviate/proto/v1/v6300/v1/base_search_pb2.py | 60 +++++++++----------
 .../proto/v1/v6300/v1/base_search_pb2.pyi     |  6 +-
 7 files changed, 103 insertions(+), 96 deletions(-)

diff --git a/weaviate/collections/grpc/shared.py b/weaviate/collections/grpc/shared.py
index 9a8949451..92af251c5 100644
--- a/weaviate/collections/grpc/shared.py
+++ b/weaviate/collections/grpc/shared.py
@@ -696,6 +696,7 @@ def _parse_hybrid(
                 if alpha is not None
                 else None,
                 alpha_param=alpha if self._weaviate_version.is_at_least(1, 36, 0) else None,
+                use_alpha_param=self._weaviate_version.is_at_least(1, 36, 0),
                 fusion_type=(
                     cast(
                         base_search_pb2.Hybrid.FusionType,
diff --git a/weaviate/proto/v1/v4216/v1/base_search_pb2.py b/weaviate/proto/v1/v4216/v1/base_search_pb2.py
index f263cb10c..3abd15ebf 100644
--- a/weaviate/proto/v1/v4216/v1/base_search_pb2.py
+++ b/weaviate/proto/v1/v4216/v1/base_search_pb2.py
@@ -14,7 +14,7 @@
 from weaviate.proto.v1.v4216.v1 import base_pb2 as v1_dot_base__pb2
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\xfe\x04\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\x11\n\x05\x61lpha\x18\x04 \x01(\x02\x42\x02\x18\x01\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x18\n\x0b\x61lpha_param\x18\x0c \x01(\x02H\x02\x88\x01\x01\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operatorB\x0e\n\x0c_alpha_param\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\x97\x05\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\x11\n\x05\x61lpha\x18\x04 \x01(\x02\x42\x02\x18\x01\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x18\n\x0b\x61lpha_param\x18\x0c \x01(\x02H\x02\x88\x01\x01\x12\x17\n\x0fuse_alpha_param\x18\r \x01(\x08\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operatorB\x0e\n\x0c_alpha_param\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -58,8 +58,8 @@
   _NEARTHERMALSEARCH.fields_by_name['target_vectors']._serialized_options = b'\030\001'
   _NEARIMUSEARCH.fields_by_name['target_vectors']._options = None
   _NEARIMUSEARCH.fields_by_name['target_vectors']._serialized_options = b'\030\001'
-  _globals['_COMBINATIONMETHOD']._serialized_start=3383
-  _globals['_COMBINATIONMETHOD']._serialized_end=3621
+  _globals['_COMBINATIONMETHOD']._serialized_start=3408
+  _globals['_COMBINATIONMETHOD']._serialized_end=3646
   _globals['_WEIGHTSFORTARGET']._serialized_start=52
   _globals['_WEIGHTSFORTARGET']._serialized_end=102
   _globals['_TARGETS']._serialized_start=105
@@ -71,31 +71,31 @@
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_start=484
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_end=555
   _globals['_HYBRID']._serialized_start=586
-  _globals['_HYBRID']._serialized_end=1224
-  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1073
-  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1170
-  _globals['_NEARVECTOR']._serialized_start=1227
-  _globals['_NEARVECTOR']._serialized_end=1656
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1575
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1629
-  _globals['_NEAROBJECT']._serialized_start=1659
-  _globals['_NEAROBJECT']._serialized_end=1824
-  _globals['_NEARTEXTSEARCH']._serialized_start=1827
-  _globals['_NEARTEXTSEARCH']._serialized_end=2195
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2088
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2142
-  _globals['_NEARIMAGESEARCH']._serialized_start=2198
-  _globals['_NEARIMAGESEARCH']._serialized_end=2371
-  _globals['_NEARAUDIOSEARCH']._serialized_start=2374
-  _globals['_NEARAUDIOSEARCH']._serialized_end=2547
-  _globals['_NEARVIDEOSEARCH']._serialized_start=2550
-  _globals['_NEARVIDEOSEARCH']._serialized_end=2723
-  _globals['_NEARDEPTHSEARCH']._serialized_start=2726
-  _globals['_NEARDEPTHSEARCH']._serialized_end=2899
-  _globals['_NEARTHERMALSEARCH']._serialized_start=2902
-  _globals['_NEARTHERMALSEARCH']._serialized_end=3079
-  _globals['_NEARIMUSEARCH']._serialized_start=3082
-  _globals['_NEARIMUSEARCH']._serialized_end=3251
-  _globals['_BM25']._serialized_start=3253
-  _globals['_BM25']._serialized_end=3380
+  _globals['_HYBRID']._serialized_end=1249
+  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1098
+  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1195
+  _globals['_NEARVECTOR']._serialized_start=1252
+  _globals['_NEARVECTOR']._serialized_end=1681
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1600
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1654
+  _globals['_NEAROBJECT']._serialized_start=1684
+  _globals['_NEAROBJECT']._serialized_end=1849
+  _globals['_NEARTEXTSEARCH']._serialized_start=1852
+  _globals['_NEARTEXTSEARCH']._serialized_end=2220
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2113
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2167
+  _globals['_NEARIMAGESEARCH']._serialized_start=2223
+  _globals['_NEARIMAGESEARCH']._serialized_end=2396
+  _globals['_NEARAUDIOSEARCH']._serialized_start=2399
+  _globals['_NEARAUDIOSEARCH']._serialized_end=2572
+  _globals['_NEARVIDEOSEARCH']._serialized_start=2575
+  _globals['_NEARVIDEOSEARCH']._serialized_end=2748
+  _globals['_NEARDEPTHSEARCH']._serialized_start=2751
+  _globals['_NEARDEPTHSEARCH']._serialized_end=2924
+  _globals['_NEARTHERMALSEARCH']._serialized_start=2927
+  _globals['_NEARTHERMALSEARCH']._serialized_end=3104
+  _globals['_NEARIMUSEARCH']._serialized_start=3107
+  _globals['_NEARIMUSEARCH']._serialized_end=3276
+  _globals['_BM25']._serialized_start=3278
+  _globals['_BM25']._serialized_end=3405
 # @@protoc_insertion_point(module_scope)
diff --git a/weaviate/proto/v1/v4216/v1/base_search_pb2.pyi b/weaviate/proto/v1/v4216/v1/base_search_pb2.pyi
index ab8a08f38..440631c88 100644
--- a/weaviate/proto/v1/v4216/v1/base_search_pb2.pyi
+++ b/weaviate/proto/v1/v4216/v1/base_search_pb2.pyi
@@ -67,7 +67,7 @@ class SearchOperatorOptions(_message.Message):
     def __init__(self, operator: _Optional[_Union[SearchOperatorOptions.Operator, str]] = ..., minimum_or_tokens_match: _Optional[int] = ...) -> None: ...
 
 class Hybrid(_message.Message):
-    __slots__ = ["query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "alpha_param", "vector_distance", "vectors"]
+    __slots__ = ["query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "alpha_param", "use_alpha_param", "vector_distance", "vectors"]
     class FusionType(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
         __slots__ = []
         FUSION_TYPE_UNSPECIFIED: _ClassVar[Hybrid.FusionType]
@@ -88,6 +88,7 @@ class Hybrid(_message.Message):
     TARGETS_FIELD_NUMBER: _ClassVar[int]
     BM25_SEARCH_OPERATOR_FIELD_NUMBER: _ClassVar[int]
     ALPHA_PARAM_FIELD_NUMBER: _ClassVar[int]
+    USE_ALPHA_PARAM_FIELD_NUMBER: _ClassVar[int]
     VECTOR_DISTANCE_FIELD_NUMBER: _ClassVar[int]
     VECTORS_FIELD_NUMBER: _ClassVar[int]
     query: str
@@ -102,9 +103,10 @@ class Hybrid(_message.Message):
     targets: Targets
     bm25_search_operator: SearchOperatorOptions
     alpha_param: float
+    use_alpha_param: bool
     vector_distance: float
     vectors: _containers.RepeatedCompositeFieldContainer[_base_pb2.Vectors]
-    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., alpha_param: _Optional[float] = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
+    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., alpha_param: _Optional[float] = ..., use_alpha_param: bool = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
 
 class NearVector(_message.Message):
     __slots__ = ["vector", "certainty", "distance", "vector_bytes", "target_vectors", "targets", "vector_per_target", "vector_for_targets", "vectors"]
diff --git a/weaviate/proto/v1/v5261/v1/base_search_pb2.py b/weaviate/proto/v1/v5261/v1/base_search_pb2.py
index cae3e4745..2658b6cc5 100644
--- a/weaviate/proto/v1/v5261/v1/base_search_pb2.py
+++ b/weaviate/proto/v1/v5261/v1/base_search_pb2.py
@@ -15,7 +15,7 @@
 from weaviate.proto.v1.v5261.v1 import base_pb2 as v1_dot_base__pb2
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\xfe\x04\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\x11\n\x05\x61lpha\x18\x04 \x01(\x02\x42\x02\x18\x01\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x18\n\x0b\x61lpha_param\x18\x0c \x01(\x02H\x02\x88\x01\x01\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operatorB\x0e\n\x0c_alpha_param\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\x97\x05\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\x11\n\x05\x61lpha\x18\x04 \x01(\x02\x42\x02\x18\x01\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x18\n\x0b\x61lpha_param\x18\x0c \x01(\x02H\x02\x88\x01\x01\x12\x17\n\x0fuse_alpha_param\x18\r \x01(\x08\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operatorB\x0e\n\x0c_alpha_param\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -59,8 +59,8 @@
   _globals['_NEARTHERMALSEARCH'].fields_by_name['target_vectors']._serialized_options = b'\030\001'
   _globals['_NEARIMUSEARCH'].fields_by_name['target_vectors']._loaded_options = None
   _globals['_NEARIMUSEARCH'].fields_by_name['target_vectors']._serialized_options = b'\030\001'
-  _globals['_COMBINATIONMETHOD']._serialized_start=3383
-  _globals['_COMBINATIONMETHOD']._serialized_end=3621
+  _globals['_COMBINATIONMETHOD']._serialized_start=3408
+  _globals['_COMBINATIONMETHOD']._serialized_end=3646
   _globals['_WEIGHTSFORTARGET']._serialized_start=52
   _globals['_WEIGHTSFORTARGET']._serialized_end=102
   _globals['_TARGETS']._serialized_start=105
@@ -72,31 +72,31 @@
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_start=484
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_end=555
   _globals['_HYBRID']._serialized_start=586
-  _globals['_HYBRID']._serialized_end=1224
-  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1073
-  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1170
-  _globals['_NEARVECTOR']._serialized_start=1227
-  _globals['_NEARVECTOR']._serialized_end=1656
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1575
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1629
-  _globals['_NEAROBJECT']._serialized_start=1659
-  _globals['_NEAROBJECT']._serialized_end=1824
-  _globals['_NEARTEXTSEARCH']._serialized_start=1827
-  _globals['_NEARTEXTSEARCH']._serialized_end=2195
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2088
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2142
-  _globals['_NEARIMAGESEARCH']._serialized_start=2198
-  _globals['_NEARIMAGESEARCH']._serialized_end=2371
-  _globals['_NEARAUDIOSEARCH']._serialized_start=2374
-  _globals['_NEARAUDIOSEARCH']._serialized_end=2547
-  _globals['_NEARVIDEOSEARCH']._serialized_start=2550
-  _globals['_NEARVIDEOSEARCH']._serialized_end=2723
-  _globals['_NEARDEPTHSEARCH']._serialized_start=2726
-  _globals['_NEARDEPTHSEARCH']._serialized_end=2899
-  _globals['_NEARTHERMALSEARCH']._serialized_start=2902
-  _globals['_NEARTHERMALSEARCH']._serialized_end=3079
-  _globals['_NEARIMUSEARCH']._serialized_start=3082
-  _globals['_NEARIMUSEARCH']._serialized_end=3251
-  _globals['_BM25']._serialized_start=3253
-  _globals['_BM25']._serialized_end=3380
+  _globals['_HYBRID']._serialized_end=1249
+  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1098
+  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1195
+  _globals['_NEARVECTOR']._serialized_start=1252
+  _globals['_NEARVECTOR']._serialized_end=1681
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1600
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1654
+  _globals['_NEAROBJECT']._serialized_start=1684
+  _globals['_NEAROBJECT']._serialized_end=1849
+  _globals['_NEARTEXTSEARCH']._serialized_start=1852
+  _globals['_NEARTEXTSEARCH']._serialized_end=2220
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2113
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2167
+  _globals['_NEARIMAGESEARCH']._serialized_start=2223
+  _globals['_NEARIMAGESEARCH']._serialized_end=2396
+  _globals['_NEARAUDIOSEARCH']._serialized_start=2399
+  _globals['_NEARAUDIOSEARCH']._serialized_end=2572
+  _globals['_NEARVIDEOSEARCH']._serialized_start=2575
+  _globals['_NEARVIDEOSEARCH']._serialized_end=2748
+  _globals['_NEARDEPTHSEARCH']._serialized_start=2751
+  _globals['_NEARDEPTHSEARCH']._serialized_end=2924
+  _globals['_NEARTHERMALSEARCH']._serialized_start=2927
+  _globals['_NEARTHERMALSEARCH']._serialized_end=3104
+  _globals['_NEARIMUSEARCH']._serialized_start=3107
+  _globals['_NEARIMUSEARCH']._serialized_end=3276
+  _globals['_BM25']._serialized_start=3278
+  _globals['_BM25']._serialized_end=3405
 # @@protoc_insertion_point(module_scope)
diff --git a/weaviate/proto/v1/v5261/v1/base_search_pb2.pyi b/weaviate/proto/v1/v5261/v1/base_search_pb2.pyi
index ca6f234bc..b89f04bfe 100644
--- a/weaviate/proto/v1/v5261/v1/base_search_pb2.pyi
+++ b/weaviate/proto/v1/v5261/v1/base_search_pb2.pyi
@@ -67,7 +67,7 @@ class SearchOperatorOptions(_message.Message):
     def __init__(self, operator: _Optional[_Union[SearchOperatorOptions.Operator, str]] = ..., minimum_or_tokens_match: _Optional[int] = ...) -> None: ...
 
 class Hybrid(_message.Message):
-    __slots__ = ("query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "alpha_param", "vector_distance", "vectors")
+    __slots__ = ("query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "alpha_param", "use_alpha_param", "vector_distance", "vectors")
     class FusionType(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
         __slots__ = ()
         FUSION_TYPE_UNSPECIFIED: _ClassVar[Hybrid.FusionType]
@@ -88,6 +88,7 @@ class Hybrid(_message.Message):
     TARGETS_FIELD_NUMBER: _ClassVar[int]
     BM25_SEARCH_OPERATOR_FIELD_NUMBER: _ClassVar[int]
     ALPHA_PARAM_FIELD_NUMBER: _ClassVar[int]
+    USE_ALPHA_PARAM_FIELD_NUMBER: _ClassVar[int]
     VECTOR_DISTANCE_FIELD_NUMBER: _ClassVar[int]
     VECTORS_FIELD_NUMBER: _ClassVar[int]
     query: str
@@ -102,9 +103,10 @@ class Hybrid(_message.Message):
     targets: Targets
     bm25_search_operator: SearchOperatorOptions
     alpha_param: float
+    use_alpha_param: bool
     vector_distance: float
     vectors: _containers.RepeatedCompositeFieldContainer[_base_pb2.Vectors]
-    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., alpha_param: _Optional[float] = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
+    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., alpha_param: _Optional[float] = ..., use_alpha_param: bool = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
 
 class NearVector(_message.Message):
     __slots__ = ("vector", "certainty", "distance", "vector_bytes", "target_vectors", "targets", "vector_per_target", "vector_for_targets", "vectors")
diff --git a/weaviate/proto/v1/v6300/v1/base_search_pb2.py b/weaviate/proto/v1/v6300/v1/base_search_pb2.py
index bc7884f55..9a19ea0ec 100644
--- a/weaviate/proto/v1/v6300/v1/base_search_pb2.py
+++ b/weaviate/proto/v1/v6300/v1/base_search_pb2.py
@@ -25,7 +25,7 @@
 from weaviate.proto.v1.v6300.v1 import base_pb2 as v1_dot_base__pb2
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\xfe\x04\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\x11\n\x05\x61lpha\x18\x04 \x01(\x02\x42\x02\x18\x01\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x18\n\x0b\x61lpha_param\x18\x0c \x01(\x02H\x02\x88\x01\x01\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operatorB\x0e\n\x0c_alpha_param\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x14v1/base_search.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\"2\n\x10WeightsForTarget\x12\x0e\n\x06target\x18\x01 \x01(\t\x12\x0e\n\x06weight\x18\x02 \x01(\x02\"\x98\x01\n\x07Targets\x12\x16\n\x0etarget_vectors\x18\x01 \x03(\t\x12\x33\n\x0b\x63ombination\x18\x02 \x01(\x0e\x32\x1e.weaviate.v1.CombinationMethod\x12:\n\x13weights_for_targets\x18\x04 \x03(\x0b\x32\x1d.weaviate.v1.WeightsForTargetJ\x04\x08\x03\x10\x04\"`\n\x0fVectorForTarget\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x18\n\x0cvector_bytes\x18\x02 \x01(\x0c\x42\x02\x18\x01\x12%\n\x07vectors\x18\x03 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"\xe1\x01\n\x15SearchOperatorOptions\x12=\n\x08operator\x18\x01 \x01(\x0e\x32+.weaviate.v1.SearchOperatorOptions.Operator\x12$\n\x17minimum_or_tokens_match\x18\x02 \x01(\x05H\x00\x88\x01\x01\"G\n\x08Operator\x12\x18\n\x14OPERATOR_UNSPECIFIED\x10\x00\x12\x0f\n\x0bOPERATOR_OR\x10\x01\x12\x10\n\x0cOPERATOR_AND\x10\x02\x42\x1a\n\x18_minimum_or_tokens_match\"\x97\x05\n\x06Hybrid\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12\x12\n\x06vector\x18\x03 \x03(\x02\x42\x02\x18\x01\x12\x11\n\x05\x61lpha\x18\x04 \x01(\x02\x42\x02\x18\x01\x12\x33\n\x0b\x66usion_type\x18\x05 \x01(\x0e\x32\x1e.weaviate.v1.Hybrid.FusionType\x12\x18\n\x0cvector_bytes\x18\x06 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x07 \x03(\tB\x02\x18\x01\x12.\n\tnear_text\x18\x08 \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearch\x12,\n\x0bnear_vector\x18\t \x01(\x0b\x32\x17.weaviate.v1.NearVector\x12%\n\x07targets\x18\n \x01(\x0b\x32\x14.weaviate.v1.Targets\x12\x45\n\x14\x62m25_search_operator\x18\x0b \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x01\x88\x01\x01\x12\x18\n\x0b\x61lpha_param\x18\x0c \x01(\x02H\x02\x88\x01\x01\x12\x17\n\x0fuse_alpha_param\x18\r \x01(\x08\x12\x19\n\x0fvector_distance\x18\x14 \x01(\x02H\x00\x12%\n\x07vectors\x18\x15 \x03(\x0b\x32\x14.weaviate.v1.Vectors\"a\n\nFusionType\x12\x1b\n\x17\x46USION_TYPE_UNSPECIFIED\x10\x00\x12\x16\n\x12\x46USION_TYPE_RANKED\x10\x01\x12\x1e\n\x1a\x46USION_TYPE_RELATIVE_SCORE\x10\x02\x42\x0b\n\tthresholdB\x17\n\x15_bm25_search_operatorB\x0e\n\x0c_alpha_param\"\xad\x03\n\nNearVector\x12\x12\n\x06vector\x18\x01 \x03(\x02\x42\x02\x18\x01\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x18\n\x0cvector_bytes\x18\x04 \x01(\x0c\x42\x02\x18\x01\x12\x1a\n\x0etarget_vectors\x18\x05 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x06 \x01(\x0b\x32\x14.weaviate.v1.Targets\x12K\n\x11vector_per_target\x18\x07 \x03(\x0b\x32,.weaviate.v1.NearVector.VectorPerTargetEntryB\x02\x18\x01\x12\x38\n\x12vector_for_targets\x18\x08 \x03(\x0b\x32\x1c.weaviate.v1.VectorForTarget\x12%\n\x07vectors\x18\t \x03(\x0b\x32\x14.weaviate.v1.Vectors\x1a\x36\n\x14VectorPerTargetEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\x0c:\x02\x38\x01\x42\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa5\x01\n\nNearObject\x12\n\n\x02id\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xf0\x02\n\x0eNearTextSearch\x12\r\n\x05query\x18\x01 \x03(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x36\n\x07move_to\x18\x04 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x02\x88\x01\x01\x12\x38\n\tmove_away\x18\x05 \x01(\x0b\x32 .weaviate.v1.NearTextSearch.MoveH\x03\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x06 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x07 \x01(\x0b\x32\x14.weaviate.v1.Targets\x1a\x36\n\x04Move\x12\r\n\x05\x66orce\x18\x01 \x01(\x02\x12\x10\n\x08\x63oncepts\x18\x02 \x03(\t\x12\r\n\x05uuids\x18\x03 \x03(\tB\x0c\n\n_certaintyB\x0b\n\t_distanceB\n\n\x08_move_toB\x0c\n\n_move_away\"\xad\x01\n\x0fNearImageSearch\x12\r\n\x05image\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearAudioSearch\x12\r\n\x05\x61udio\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearVideoSearch\x12\r\n\x05video\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xad\x01\n\x0fNearDepthSearch\x12\r\n\x05\x64\x65pth\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xb1\x01\n\x11NearThermalSearch\x12\x0f\n\x07thermal\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\xa9\x01\n\rNearIMUSearch\x12\x0b\n\x03imu\x18\x01 \x01(\t\x12\x16\n\tcertainty\x18\x02 \x01(\x01H\x00\x88\x01\x01\x12\x15\n\x08\x64istance\x18\x03 \x01(\x01H\x01\x88\x01\x01\x12\x1a\n\x0etarget_vectors\x18\x04 \x03(\tB\x02\x18\x01\x12%\n\x07targets\x18\x05 \x01(\x0b\x32\x14.weaviate.v1.TargetsB\x0c\n\n_certaintyB\x0b\n\t_distance\"\x7f\n\x04\x42M25\x12\r\n\x05query\x18\x01 \x01(\t\x12\x12\n\nproperties\x18\x02 \x03(\t\x12@\n\x0fsearch_operator\x18\x03 \x01(\x0b\x32\".weaviate.v1.SearchOperatorOptionsH\x00\x88\x01\x01\x42\x12\n\x10_search_operator*\xee\x01\n\x11\x43ombinationMethod\x12\"\n\x1e\x43OMBINATION_METHOD_UNSPECIFIED\x10\x00\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_SUM\x10\x01\x12\x1f\n\x1b\x43OMBINATION_METHOD_TYPE_MIN\x10\x02\x12#\n\x1f\x43OMBINATION_METHOD_TYPE_AVERAGE\x10\x03\x12*\n&COMBINATION_METHOD_TYPE_RELATIVE_SCORE\x10\x04\x12\"\n\x1e\x43OMBINATION_METHOD_TYPE_MANUAL\x10\x05\x42t\n#io.weaviate.client.grpc.protocol.v1B\x17WeaviateProtoBaseSearchZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -69,8 +69,8 @@
   _globals['_NEARTHERMALSEARCH'].fields_by_name['target_vectors']._serialized_options = b'\030\001'
   _globals['_NEARIMUSEARCH'].fields_by_name['target_vectors']._loaded_options = None
   _globals['_NEARIMUSEARCH'].fields_by_name['target_vectors']._serialized_options = b'\030\001'
-  _globals['_COMBINATIONMETHOD']._serialized_start=3383
-  _globals['_COMBINATIONMETHOD']._serialized_end=3621
+  _globals['_COMBINATIONMETHOD']._serialized_start=3408
+  _globals['_COMBINATIONMETHOD']._serialized_end=3646
   _globals['_WEIGHTSFORTARGET']._serialized_start=52
   _globals['_WEIGHTSFORTARGET']._serialized_end=102
   _globals['_TARGETS']._serialized_start=105
@@ -82,31 +82,31 @@
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_start=484
   _globals['_SEARCHOPERATOROPTIONS_OPERATOR']._serialized_end=555
   _globals['_HYBRID']._serialized_start=586
-  _globals['_HYBRID']._serialized_end=1224
-  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1073
-  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1170
-  _globals['_NEARVECTOR']._serialized_start=1227
-  _globals['_NEARVECTOR']._serialized_end=1656
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1575
-  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1629
-  _globals['_NEAROBJECT']._serialized_start=1659
-  _globals['_NEAROBJECT']._serialized_end=1824
-  _globals['_NEARTEXTSEARCH']._serialized_start=1827
-  _globals['_NEARTEXTSEARCH']._serialized_end=2195
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2088
-  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2142
-  _globals['_NEARIMAGESEARCH']._serialized_start=2198
-  _globals['_NEARIMAGESEARCH']._serialized_end=2371
-  _globals['_NEARAUDIOSEARCH']._serialized_start=2374
-  _globals['_NEARAUDIOSEARCH']._serialized_end=2547
-  _globals['_NEARVIDEOSEARCH']._serialized_start=2550
-  _globals['_NEARVIDEOSEARCH']._serialized_end=2723
-  _globals['_NEARDEPTHSEARCH']._serialized_start=2726
-  _globals['_NEARDEPTHSEARCH']._serialized_end=2899
-  _globals['_NEARTHERMALSEARCH']._serialized_start=2902
-  _globals['_NEARTHERMALSEARCH']._serialized_end=3079
-  _globals['_NEARIMUSEARCH']._serialized_start=3082
-  _globals['_NEARIMUSEARCH']._serialized_end=3251
-  _globals['_BM25']._serialized_start=3253
-  _globals['_BM25']._serialized_end=3380
+  _globals['_HYBRID']._serialized_end=1249
+  _globals['_HYBRID_FUSIONTYPE']._serialized_start=1098
+  _globals['_HYBRID_FUSIONTYPE']._serialized_end=1195
+  _globals['_NEARVECTOR']._serialized_start=1252
+  _globals['_NEARVECTOR']._serialized_end=1681
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_start=1600
+  _globals['_NEARVECTOR_VECTORPERTARGETENTRY']._serialized_end=1654
+  _globals['_NEAROBJECT']._serialized_start=1684
+  _globals['_NEAROBJECT']._serialized_end=1849
+  _globals['_NEARTEXTSEARCH']._serialized_start=1852
+  _globals['_NEARTEXTSEARCH']._serialized_end=2220
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_start=2113
+  _globals['_NEARTEXTSEARCH_MOVE']._serialized_end=2167
+  _globals['_NEARIMAGESEARCH']._serialized_start=2223
+  _globals['_NEARIMAGESEARCH']._serialized_end=2396
+  _globals['_NEARAUDIOSEARCH']._serialized_start=2399
+  _globals['_NEARAUDIOSEARCH']._serialized_end=2572
+  _globals['_NEARVIDEOSEARCH']._serialized_start=2575
+  _globals['_NEARVIDEOSEARCH']._serialized_end=2748
+  _globals['_NEARDEPTHSEARCH']._serialized_start=2751
+  _globals['_NEARDEPTHSEARCH']._serialized_end=2924
+  _globals['_NEARTHERMALSEARCH']._serialized_start=2927
+  _globals['_NEARTHERMALSEARCH']._serialized_end=3104
+  _globals['_NEARIMUSEARCH']._serialized_start=3107
+  _globals['_NEARIMUSEARCH']._serialized_end=3276
+  _globals['_BM25']._serialized_start=3278
+  _globals['_BM25']._serialized_end=3405
 # @@protoc_insertion_point(module_scope)
diff --git a/weaviate/proto/v1/v6300/v1/base_search_pb2.pyi b/weaviate/proto/v1/v6300/v1/base_search_pb2.pyi
index 9ad165a42..67a53e25a 100644
--- a/weaviate/proto/v1/v6300/v1/base_search_pb2.pyi
+++ b/weaviate/proto/v1/v6300/v1/base_search_pb2.pyi
@@ -68,7 +68,7 @@ class SearchOperatorOptions(_message.Message):
     def __init__(self, operator: _Optional[_Union[SearchOperatorOptions.Operator, str]] = ..., minimum_or_tokens_match: _Optional[int] = ...) -> None: ...
 
 class Hybrid(_message.Message):
-    __slots__ = ("query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "alpha_param", "vector_distance", "vectors")
+    __slots__ = ("query", "properties", "vector", "alpha", "fusion_type", "vector_bytes", "target_vectors", "near_text", "near_vector", "targets", "bm25_search_operator", "alpha_param", "use_alpha_param", "vector_distance", "vectors")
     class FusionType(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
         __slots__ = ()
         FUSION_TYPE_UNSPECIFIED: _ClassVar[Hybrid.FusionType]
@@ -89,6 +89,7 @@ class Hybrid(_message.Message):
     TARGETS_FIELD_NUMBER: _ClassVar[int]
     BM25_SEARCH_OPERATOR_FIELD_NUMBER: _ClassVar[int]
     ALPHA_PARAM_FIELD_NUMBER: _ClassVar[int]
+    USE_ALPHA_PARAM_FIELD_NUMBER: _ClassVar[int]
     VECTOR_DISTANCE_FIELD_NUMBER: _ClassVar[int]
     VECTORS_FIELD_NUMBER: _ClassVar[int]
     query: str
@@ -103,9 +104,10 @@ class Hybrid(_message.Message):
     targets: Targets
     bm25_search_operator: SearchOperatorOptions
     alpha_param: float
+    use_alpha_param: bool
     vector_distance: float
     vectors: _containers.RepeatedCompositeFieldContainer[_base_pb2.Vectors]
-    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., alpha_param: _Optional[float] = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
+    def __init__(self, query: _Optional[str] = ..., properties: _Optional[_Iterable[str]] = ..., vector: _Optional[_Iterable[float]] = ..., alpha: _Optional[float] = ..., fusion_type: _Optional[_Union[Hybrid.FusionType, str]] = ..., vector_bytes: _Optional[bytes] = ..., target_vectors: _Optional[_Iterable[str]] = ..., near_text: _Optional[_Union[NearTextSearch, _Mapping]] = ..., near_vector: _Optional[_Union[NearVector, _Mapping]] = ..., targets: _Optional[_Union[Targets, _Mapping]] = ..., bm25_search_operator: _Optional[_Union[SearchOperatorOptions, _Mapping]] = ..., alpha_param: _Optional[float] = ..., use_alpha_param: bool = ..., vector_distance: _Optional[float] = ..., vectors: _Optional[_Iterable[_Union[_base_pb2.Vectors, _Mapping]]] = ...) -> None: ...
 
 class NearVector(_message.Message):
     __slots__ = ("vector", "certainty", "distance", "vector_bytes", "target_vectors", "targets", "vector_per_target", "vector_for_targets", "vectors")

From 8e5d54d7cd09fc20bc4f1f87b7678b1cdafaa05c Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 12:14:23 +0000
Subject: [PATCH 04/99] Change formatting

---
 weaviate/collections/grpc/shared.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/weaviate/collections/grpc/shared.py b/weaviate/collections/grpc/shared.py
index 92af251c5..c5131c51e 100644
--- a/weaviate/collections/grpc/shared.py
+++ b/weaviate/collections/grpc/shared.py
@@ -692,7 +692,7 @@ def _parse_hybrid(
                 query=query,
                 alpha=None
                 if self._weaviate_version.is_at_least(1, 36, 0)
-                else (alpha)
+                else alpha
                 if alpha is not None
                 else None,
                 alpha_param=alpha if self._weaviate_version.is_at_least(1, 36, 0) else None,

From 367ce7965ad5785f54ac647dc16a23fe5c4894cb Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 12:15:54 +0000
Subject: [PATCH 05/99] Tidy version check code

---
 weaviate/collections/grpc/shared.py | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

diff --git a/weaviate/collections/grpc/shared.py b/weaviate/collections/grpc/shared.py
index c5131c51e..ff538020c 100644
--- a/weaviate/collections/grpc/shared.py
+++ b/weaviate/collections/grpc/shared.py
@@ -686,17 +686,14 @@ def _parse_hybrid(
             else:
                 vector_bytes = vector_bytes_tmp
 
+        is_1_36 = self._weaviate_version.is_at_least(1, 36, 0)
         return (
             base_search_pb2.Hybrid(
                 properties=properties,
                 query=query,
-                alpha=None
-                if self._weaviate_version.is_at_least(1, 36, 0)
-                else alpha
-                if alpha is not None
-                else None,
-                alpha_param=alpha if self._weaviate_version.is_at_least(1, 36, 0) else None,
-                use_alpha_param=self._weaviate_version.is_at_least(1, 36, 0),
+                alpha=None if is_1_36 else (alpha if alpha is not None else None),
+                alpha_param=alpha if is_1_36 else None,
+                use_alpha_param=is_1_36,
                 fusion_type=(
                     cast(
                         base_search_pb2.Hybrid.FusionType,

From 81414a11736c60f532fee516e9ad7a4999092265 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 12:17:52 +0000
Subject: [PATCH 06/99] Parse correct default for BC if server < 1.36

---
 weaviate/collections/grpc/shared.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/weaviate/collections/grpc/shared.py b/weaviate/collections/grpc/shared.py
index ff538020c..6a08e9406 100644
--- a/weaviate/collections/grpc/shared.py
+++ b/weaviate/collections/grpc/shared.py
@@ -691,7 +691,7 @@ def _parse_hybrid(
             base_search_pb2.Hybrid(
                 properties=properties,
                 query=query,
-                alpha=None if is_1_36 else (alpha if alpha is not None else None),
+                alpha=None if is_1_36 else (alpha if alpha is not None else 0.7),
                 alpha_param=alpha if is_1_36 else None,
                 use_alpha_param=is_1_36,
                 fusion_type=(

From a13da2373b9f26c5bb4bb9d136e5c53a81df1251 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 12:19:47 +0000
Subject: [PATCH 07/99] Update CI image

---
 .github/workflows/main.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index 1fa7afb20..1edbe48cf 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -28,7 +28,7 @@ env:
   WEAVIATE_134: 1.34.5
   WEAVIATE_135: 1.35.0
   WEAVIATE_136: 1.36.0
-  WEAVIATE_137: 1.37.0-dev-29d5c87.amd64
+  WEAVIATE_137: 1.37.0-dev-8ff93a2.amd64
 
 
 jobs:

From 7abe7f534e5976c28c6fc489111d2080489612e0 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 15:00:29 +0000
Subject: [PATCH 08/99] Fix wrong version comparison

---
 weaviate/collections/grpc/shared.py | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/weaviate/collections/grpc/shared.py b/weaviate/collections/grpc/shared.py
index 6a08e9406..cc3be7ec1 100644
--- a/weaviate/collections/grpc/shared.py
+++ b/weaviate/collections/grpc/shared.py
@@ -686,14 +686,14 @@ def _parse_hybrid(
             else:
                 vector_bytes = vector_bytes_tmp
 
-        is_1_36 = self._weaviate_version.is_at_least(1, 36, 0)
+        is_1_37 = self._weaviate_version.is_at_least(1, 37, 0)
         return (
             base_search_pb2.Hybrid(
                 properties=properties,
                 query=query,
-                alpha=None if is_1_36 else (alpha if alpha is not None else 0.7),
-                alpha_param=alpha if is_1_36 else None,
-                use_alpha_param=is_1_36,
+                alpha=None if is_1_37 else (alpha if alpha is not None else 0.7),
+                alpha_param=alpha if is_1_37 else None,
+                use_alpha_param=is_1_37,
                 fusion_type=(
                     cast(
                         base_search_pb2.Hybrid.FusionType,

From b5f5e559e73dbefcc5b9487aa045ade07eee0062 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 15:12:09 +0000
Subject: [PATCH 09/99] Fix typo in ci

---
 .github/workflows/main.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index 1edbe48cf..cef92023c 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -310,7 +310,7 @@ jobs:
           $WEAVIATE_133,
           $WEAVIATE_134,
           $WEAVIATE_135,
-          $WEAVIATE_136
+          $WEAVIATE_136,
           $WEAVIATE_137
         ]
     steps:

From 47841d4ed251bfd6b35da5a4030993d697a2598d Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 16:13:15 +0000
Subject: [PATCH 10/99] Update CI image

---
 .github/workflows/main.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index cef92023c..43253fbb4 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -28,7 +28,7 @@ env:
   WEAVIATE_134: 1.34.5
   WEAVIATE_135: 1.35.0
   WEAVIATE_136: 1.36.0
-  WEAVIATE_137: 1.37.0-dev-8ff93a2.amd64
+  WEAVIATE_137: 1.37.0-dev-4e2d51d.amd64
 
 
 jobs:

From e305b334a2e9b05908801d79c31c1a11f68785be Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 16 Mar 2026 16:14:59 +0000
Subject: [PATCH 11/99] Remove client-side default from aggregate queries

---
 weaviate/collections/aggregations/hybrid/async_.pyi  | 6 +++---
 weaviate/collections/aggregations/hybrid/executor.py | 8 ++++----
 weaviate/collections/aggregations/hybrid/sync.pyi    | 6 +++---
 3 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/weaviate/collections/aggregations/hybrid/async_.pyi b/weaviate/collections/aggregations/hybrid/async_.pyi
index 8f217a5d4..f1e4b72ed 100644
--- a/weaviate/collections/aggregations/hybrid/async_.pyi
+++ b/weaviate/collections/aggregations/hybrid/async_.pyi
@@ -19,7 +19,7 @@ class _HybridAsync(_HybridExecutor[ConnectionAsync]):
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,
@@ -36,7 +36,7 @@ class _HybridAsync(_HybridExecutor[ConnectionAsync]):
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,
@@ -53,7 +53,7 @@ class _HybridAsync(_HybridExecutor[ConnectionAsync]):
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,
diff --git a/weaviate/collections/aggregations/hybrid/executor.py b/weaviate/collections/aggregations/hybrid/executor.py
index fae005c9c..c2c87a68f 100644
--- a/weaviate/collections/aggregations/hybrid/executor.py
+++ b/weaviate/collections/aggregations/hybrid/executor.py
@@ -22,7 +22,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,
@@ -40,7 +40,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,
@@ -58,7 +58,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,
@@ -75,7 +75,7 @@ def hybrid(
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,
diff --git a/weaviate/collections/aggregations/hybrid/sync.pyi b/weaviate/collections/aggregations/hybrid/sync.pyi
index 81d632f2e..8656fb319 100644
--- a/weaviate/collections/aggregations/hybrid/sync.pyi
+++ b/weaviate/collections/aggregations/hybrid/sync.pyi
@@ -19,7 +19,7 @@ class _Hybrid(_HybridExecutor[ConnectionSync]):
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,
@@ -36,7 +36,7 @@ class _Hybrid(_HybridExecutor[ConnectionSync]):
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,
@@ -53,7 +53,7 @@ class _Hybrid(_HybridExecutor[ConnectionSync]):
         self,
         query: Optional[str],
         *,
-        alpha: NUMBER = 0.7,
+        alpha: Optional[NUMBER] = None,
         vector: Optional[List[float]] = None,
         query_properties: Optional[List[str]] = None,
         object_limit: Optional[int] = None,

From 8520a1c4f418f767732404fc4c6c44e9eef722df Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 11:58:41 +0000
Subject: [PATCH 12/99] Update ver check and CI tags

---
 .github/workflows/main.yaml         | 15 +++++++--------
 weaviate/collections/grpc/shared.py | 10 ++++++----
 2 files changed, 13 insertions(+), 12 deletions(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index 43253fbb4..d884891c0 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -22,14 +22,13 @@ env:
   WEAVIATE_128: 1.28.16
   WEAVIATE_129: 1.29.11
   WEAVIATE_130: 1.30.22
-  WEAVIATE_131: 1.31.20
-  WEAVIATE_132: 1.32.23
-  WEAVIATE_133: 1.33.10
-  WEAVIATE_134: 1.34.5
-  WEAVIATE_135: 1.35.0
-  WEAVIATE_136: 1.36.0
-  WEAVIATE_137: 1.37.0-dev-4e2d51d.amd64
-
+  WEAVIATE_131: 1.31.22
+  WEAVIATE_132: 1.32.27
+  WEAVIATE_133: 1.33.18
+  WEAVIATE_134: 1.34.19
+  WEAVIATE_135: 1.35.15
+  WEAVIATE_136: 1.36.6-1abf310
+  WEAVIATE_137: 1.37.0-dev-7eb294e
 
 jobs:
   lint-and-format:
diff --git a/weaviate/collections/grpc/shared.py b/weaviate/collections/grpc/shared.py
index cc3be7ec1..714db67cb 100644
--- a/weaviate/collections/grpc/shared.py
+++ b/weaviate/collections/grpc/shared.py
@@ -686,14 +686,16 @@ def _parse_hybrid(
             else:
                 vector_bytes = vector_bytes_tmp
 
-        is_1_37 = self._weaviate_version.is_at_least(1, 37, 0)
+        use_alpha_param = self._weaviate_version.is_at_least(
+            1, 36, 6
+        )  # TODO: change to 1.36.7 once it's released
         return (
             base_search_pb2.Hybrid(
                 properties=properties,
                 query=query,
-                alpha=None if is_1_37 else (alpha if alpha is not None else 0.7),
-                alpha_param=alpha if is_1_37 else None,
-                use_alpha_param=is_1_37,
+                alpha=None if use_alpha_param else (alpha if alpha is not None else 0.7),
+                alpha_param=alpha if use_alpha_param else None,
+                use_alpha_param=use_alpha_param,
                 fusion_type=(
                     cast(
                         base_search_pb2.Hybrid.FusionType,

From ed373755688b09d105a932d15326993ac54059f0 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 12:30:15 +0000
Subject: [PATCH 13/99] Remove test of lazy loading shards

---
 integration/test_client.py | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/integration/test_client.py b/integration/test_client.py
index 23ece9885..61d3eb64a 100644
--- a/integration/test_client.py
+++ b/integration/test_client.py
@@ -383,10 +383,7 @@ def test_client_cluster_without_lazy_shard_loading(
         ]
         assert nodes[0].shards[0].vector_queue_length == 0
         assert nodes[0].shards[0].compressed is False
-        if collection._connection._weaviate_version.is_lower_than(1, 25, 0):
-            assert nodes[0].shards[0].loaded is True
-        else:
-            assert nodes[0].shards[0].loaded is False
+        assert nodes[0].shards[0].loaded is True
     finally:
         client.collections.delete(request.node.name)
 

From 14339851c0c5fa9cd3a173369aeb65ccc3ca37f0 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 13:01:58 +0000
Subject: [PATCH 14/99] Refactor client test for new server lazy shard loading

---
 integration/test_client.py | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/integration/test_client.py b/integration/test_client.py
index 61d3eb64a..4166cb726 100644
--- a/integration/test_client.py
+++ b/integration/test_client.py
@@ -355,7 +355,10 @@ def test_client_cluster_with_lazy_shard_loading(
         ]
         assert nodes[0].shards[0].vector_queue_length == 0
         assert nodes[0].shards[0].compressed is False
-        assert nodes[0].shards[0].loaded is True
+        if collection._connection._weaviate_version.is_lower_than(1, 36, 0):
+            assert nodes[0].shards[0].loaded is False
+        else:
+            assert nodes[0].shards[0].loaded is True
     finally:
         client.collections.delete(request.node.name)
 
@@ -383,7 +386,10 @@ def test_client_cluster_without_lazy_shard_loading(
         ]
         assert nodes[0].shards[0].vector_queue_length == 0
         assert nodes[0].shards[0].compressed is False
-        assert nodes[0].shards[0].loaded is True
+        if collection._connection._weaviate_version.is_lower_than(1, 25, 0):
+            assert nodes[0].shards[0].loaded is True
+        else:
+            assert nodes[0].shards[0].loaded is False
     finally:
         client.collections.delete(request.node.name)
 

From dd6c835f9d4f66409e3010840da76945ddc58e7f Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 13:18:16 +0000
Subject: [PATCH 15/99] Debug failing ci test

---
 integration/test_client.py | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/integration/test_client.py b/integration/test_client.py
index 4166cb726..23ece9885 100644
--- a/integration/test_client.py
+++ b/integration/test_client.py
@@ -355,10 +355,7 @@ def test_client_cluster_with_lazy_shard_loading(
         ]
         assert nodes[0].shards[0].vector_queue_length == 0
         assert nodes[0].shards[0].compressed is False
-        if collection._connection._weaviate_version.is_lower_than(1, 36, 0):
-            assert nodes[0].shards[0].loaded is False
-        else:
-            assert nodes[0].shards[0].loaded is True
+        assert nodes[0].shards[0].loaded is True
     finally:
         client.collections.delete(request.node.name)
 

From cf96ccb40309a1001abb6c533c6d9b8cec2e9723 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 13:41:55 +0000
Subject: [PATCH 16/99] Remove outdated lazy shard load test

---
 integration/test_client.py | 33 +--------------------------------
 1 file changed, 1 insertion(+), 32 deletions(-)

diff --git a/integration/test_client.py b/integration/test_client.py
index 23ece9885..3560cbbb7 100644
--- a/integration/test_client.py
+++ b/integration/test_client.py
@@ -334,7 +334,7 @@ def test_collection_name_capitalization(
         client.collections.delete(name_big)
 
 
-def test_client_cluster_with_lazy_shard_loading(
+def test_client_cluster_without_lazy_shard_loading(
     client: weaviate.WeaviateClient, request: SubRequest
 ) -> None:
     try:
@@ -360,37 +360,6 @@ def test_client_cluster_with_lazy_shard_loading(
         client.collections.delete(request.node.name)
 
 
-def test_client_cluster_without_lazy_shard_loading(
-    client_factory: ClientFactory, request: SubRequest
-) -> None:
-    client = client_factory(8090, 50061)
-
-    try:
-        collection = client.collections.create(
-            name=request.node.name, vectorizer_config=Configure.Vectorizer.none()
-        )
-
-        nodes = client.cluster.nodes(collection.name, output="verbose")
-        assert len(nodes) == 1
-        assert len(nodes[0].shards) == 1
-        assert nodes[0].shards[0].collection == collection.name
-        assert nodes[0].shards[0].object_count == 0
-        assert nodes[0].shards[0].vector_indexing_status in [
-            "READONLY",
-            "INDEXING",
-            "READY",
-            "LAZY_LOADING",
-        ]
-        assert nodes[0].shards[0].vector_queue_length == 0
-        assert nodes[0].shards[0].compressed is False
-        if collection._connection._weaviate_version.is_lower_than(1, 25, 0):
-            assert nodes[0].shards[0].loaded is True
-        else:
-            assert nodes[0].shards[0].loaded is False
-    finally:
-        client.collections.delete(request.node.name)
-
-
 def test_client_cluster_multitenant(client: weaviate.WeaviateClient, request: SubRequest) -> None:
     try:
         collection = client.collections.create(

From 4b669e10706fca04573f5f767d11c6e848b0b02d Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 15:21:09 +0000
Subject: [PATCH 17/99] Update CI images

---
 .github/workflows/main.yaml | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index d884891c0..7cc9934e7 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -27,8 +27,8 @@ env:
   WEAVIATE_133: 1.33.18
   WEAVIATE_134: 1.34.19
   WEAVIATE_135: 1.35.15
-  WEAVIATE_136: 1.36.6-1abf310
-  WEAVIATE_137: 1.37.0-dev-7eb294e
+  WEAVIATE_136: 1.36.6-21aaadc
+  WEAVIATE_137: 1.37.0-dev-8fb696b
 
 jobs:
   lint-and-format:

From 890a414aa43eeb355f40c7ed9402a1edd45f40cd Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 15:59:52 +0000
Subject: [PATCH 18/99] Add per-test timeouts and stack dump on timeout

---
 integration/conftest.py | 63 ++++++++++++++++++++++++++++++++++++++++-
 pytest.ini              |  4 ++-
 requirements-test.txt   |  1 +
 3 files changed, 66 insertions(+), 2 deletions(-)

diff --git a/integration/conftest.py b/integration/conftest.py
index 256517ea6..891c66307 100644
--- a/integration/conftest.py
+++ b/integration/conftest.py
@@ -1,5 +1,10 @@
 import os
 import time
+import sys
+import threading
+import traceback
+import signal
+import pytest
 from typing import (
     Any,
     AsyncGenerator,
@@ -14,7 +19,6 @@
 )
 from typing import Callable, TypeVar
 
-import pytest
 import pytest_asyncio
 from _pytest.fixtures import SubRequest
 
@@ -500,3 +504,60 @@ def retry_on_http_error(
             raise
     # This should never be reached, but satisfies the type checker
     raise last_exception  # type: ignore
+
+
+TIMEOUT_SECONDS = 30
+
+
+def dump_all_stacks():
+    """Print stack traces for every live thread."""
+    frames = sys._current_frames()
+    lines = ["\n===== DEADLOCK DETECTED — THREAD DUMP =====\n"]
+    for thread in threading.enumerate():
+        frame = frames.get(thread.ident)  # pyright: ignore
+        lines.append(f"\n--- Thread: {thread.name} (id={thread.ident}, daemon={thread.daemon}) ---")
+        if frame:
+            lines.append("".join(traceback.format_stack(frame)))
+        else:
+            lines.append("  (no frame available)\n")
+    lines.append("===========================================\n")
+    return "\n".join(lines)
+
+
+class DeadlockWatchdog:
+    def __init__(self, timeout):
+        self.timeout = timeout
+        self._timer = None
+        self._test_name = None
+
+    def start(self, test_name):
+        self._test_name = test_name
+        self._timer = threading.Timer(self.timeout, self._on_timeout)
+        self._timer.daemon = True
+        self._timer.start()
+
+    def stop(self):
+        if self._timer:
+            self._timer.cancel()
+            self._timer = None
+
+    def _on_timeout(self):
+        dump = dump_all_stacks()
+        # Write to stderr so it's always visible even if captured
+        sys.stderr.write(f"\n[WATCHDOG] Test '{self._test_name}' timed out after {self.timeout}s\n")
+        sys.stderr.write(dump)
+        sys.stderr.flush()
+        # Force-kill the process so CI doesn't hang forever
+        signal.raise_signal(signal.SIGTERM)
+
+
+_watchdog = DeadlockWatchdog(TIMEOUT_SECONDS)
+
+
+@pytest.hookimpl(hookwrapper=True)
+def pytest_runtest_call(item):
+    _watchdog.start(item.nodeid)
+    try:
+        yield
+    finally:
+        _watchdog.stop()
diff --git a/pytest.ini b/pytest.ini
index 36321aa1d..09bfb6983 100644
--- a/pytest.ini
+++ b/pytest.ini
@@ -2,4 +2,6 @@
 addopts = -m 'not profiling' --benchmark-skip -l
 markers =
     profiling: marks tests that can be profiled
-asyncio_default_fixture_loop_scope = function
\ No newline at end of file
+asyncio_default_fixture_loop_scope = function
+timeout = 600 # 10 minutes
+timeout_method = thread
\ No newline at end of file
diff --git a/requirements-test.txt b/requirements-test.txt
index 9817a6d5e..40121fcce 100644
--- a/requirements-test.txt
+++ b/requirements-test.txt
@@ -3,6 +3,7 @@ pytest-cov==6.2.1
 pytest-asyncio==1.3.0
 pytest-benchmark==5.1.0
 pytest-profiling==1.8.1
+pytest-timeout==2.4.0
 coverage==7.10.7
 pytest-xdist==3.7.0
 werkzeug==3.1.6

From 2893822041959749f6cd12554f304c114d4015e0 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 16:01:09 +0000
Subject: [PATCH 19/99] Reduce per-test timeout to 5 mins

---
 pytest.ini | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pytest.ini b/pytest.ini
index 09bfb6983..36de4d551 100644
--- a/pytest.ini
+++ b/pytest.ini
@@ -3,5 +3,5 @@ addopts = -m 'not profiling' --benchmark-skip -l
 markers =
     profiling: marks tests that can be profiled
 asyncio_default_fixture_loop_scope = function
-timeout = 600 # 10 minutes
+timeout = 300 # 5 minutes
 timeout_method = thread
\ No newline at end of file

From 6edd0e155630549f65ead7ff9a2c63baedca363d Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 16:03:32 +0000
Subject: [PATCH 20/99] Fix inc backups test

---
 integration/test_backup_v4.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/integration/test_backup_v4.py b/integration/test_backup_v4.py
index 1204a1269..6da10bc7c 100644
--- a/integration/test_backup_v4.py
+++ b/integration/test_backup_v4.py
@@ -764,7 +764,7 @@ def test_incremental_backup(client: weaviate.WeaviateClient, request: SubRequest
         backend=BACKEND,
         include_collections=["Article"],
         wait_for_completion=True,
-        incremental_backup_base_id=base_backup_id,
+        incremental_base_backup_id=base_backup_id,
     )
 
     # remove existing class

From e7cb6979581e87a5836d6e1dbc983a989eed284b Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 16:04:07 +0000
Subject: [PATCH 21/99] Add server version check for incremental backups

---
 weaviate/backup/executor.py | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/weaviate/backup/executor.py b/weaviate/backup/executor.py
index e8c52239d..09a47bd0a 100644
--- a/weaviate/backup/executor.py
+++ b/weaviate/backup/executor.py
@@ -88,6 +88,16 @@ def create(
             wait_for_completion=wait_for_completion,
         )
 
+        if (
+            incremental_base_backup_id is not None
+            and self._connection._weaviate_version.is_lower_than(1, 37, 0)
+        ):
+            raise WeaviateUnsupportedFeatureError(
+                "Incremental backups",
+                str(self._connection._weaviate_version),
+                "1.37.0",
+            )
+
         payload: dict = {
             "id": backup_id,
             "include": include_collections,

From 12d064ff617a400c3348ad8eac42980efea99597 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 16:09:03 +0000
Subject: [PATCH 22/99] Remove comment

---
 pytest.ini | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pytest.ini b/pytest.ini
index 36de4d551..8b61c6947 100644
--- a/pytest.ini
+++ b/pytest.ini
@@ -3,5 +3,5 @@ addopts = -m 'not profiling' --benchmark-skip -l
 markers =
     profiling: marks tests that can be profiled
 asyncio_default_fixture_loop_scope = function
-timeout = 300 # 5 minutes
+timeout = 300
 timeout_method = thread
\ No newline at end of file

From fe5c5223901ebcf67f7f693e72278d99a9bed5cc Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 16:34:16 +0000
Subject: [PATCH 23/99] Hard kill the process on timeout detection

---
 integration/conftest.py | 35 ++++++++++++++++++++---------------
 1 file changed, 20 insertions(+), 15 deletions(-)

diff --git a/integration/conftest.py b/integration/conftest.py
index 891c66307..73a032f57 100644
--- a/integration/conftest.py
+++ b/integration/conftest.py
@@ -3,7 +3,6 @@
 import sys
 import threading
 import traceback
-import signal
 import pytest
 from typing import (
     Any,
@@ -506,20 +505,19 @@ def retry_on_http_error(
     raise last_exception  # type: ignore
 
 
-TIMEOUT_SECONDS = 30
+TIMEOUT_SECONDS = 300
 
 
 def dump_all_stacks():
-    """Print stack traces for every live thread."""
     frames = sys._current_frames()
     lines = ["\n===== DEADLOCK DETECTED — THREAD DUMP =====\n"]
     for thread in threading.enumerate():
         frame = frames.get(thread.ident)  # pyright: ignore
-        lines.append(f"\n--- Thread: {thread.name} (id={thread.ident}, daemon={thread.daemon}) ---")
+        lines.append(f"\n--- Thread: {thread.name} (id={thread.ident}) ---")
         if frame:
             lines.append("".join(traceback.format_stack(frame)))
         else:
-            lines.append("  (no frame available)\n")
+            lines.append("  (no frame)\n")
     lines.append("===========================================\n")
     return "\n".join(lines)
 
@@ -528,10 +526,9 @@ class DeadlockWatchdog:
     def __init__(self, timeout):
         self.timeout = timeout
         self._timer = None
-        self._test_name = None
 
-    def start(self, test_name):
-        self._test_name = test_name
+    def start(self, label):
+        self._label = label
         self._timer = threading.Timer(self.timeout, self._on_timeout)
         self._timer.daemon = True
         self._timer.start()
@@ -542,22 +539,30 @@ def stop(self):
             self._timer = None
 
     def _on_timeout(self):
-        dump = dump_all_stacks()
-        # Write to stderr so it's always visible even if captured
-        sys.stderr.write(f"\n[WATCHDOG] Test '{self._test_name}' timed out after {self.timeout}s\n")
-        sys.stderr.write(dump)
+        sys.stderr.write(f"\n[WATCHDOG] Hung at: '{self._label}' after {self.timeout}s\n")
+        sys.stderr.write(dump_all_stacks())
         sys.stderr.flush()
-        # Force-kill the process so CI doesn't hang forever
-        signal.raise_signal(signal.SIGTERM)
+        os._exit(1)  # Hard kill — works reliably in xdist workers
 
 
 _watchdog = DeadlockWatchdog(TIMEOUT_SECONDS)
 
 
+# Covers setup + call + teardown
 @pytest.hookimpl(hookwrapper=True)
-def pytest_runtest_call(item):
+def pytest_runtest_protocol(item, nextitem):
     _watchdog.start(item.nodeid)
     try:
         yield
     finally:
         _watchdog.stop()
+
+
+# Separately watch session-scoped fixture setup
+@pytest.hookimpl(hookwrapper=True)
+def pytest_sessionstart(session):
+    _watchdog.start("session startup / session-scoped fixtures")
+    try:
+        yield
+    finally:
+        _watchdog.stop()

From 2d961d18842bcf678f62d666b7a2a44388ea4dbc Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 16:58:40 +0000
Subject: [PATCH 24/99] Timeout putting the sentinel to avoid deadlocking

---
 weaviate/collections/batch/async_.py | 2 +-
 weaviate/collections/batch/sync.py   | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/weaviate/collections/batch/async_.py b/weaviate/collections/batch/async_.py
index b71f8be39..b86bc80d3 100644
--- a/weaviate/collections/batch/async_.py
+++ b/weaviate/collections/batch/async_.py
@@ -254,7 +254,7 @@ async def __loop(self) -> None:
                 and not self.__is_shutting_down.is_set()
                 and not self.__is_oom.is_set()
             ):
-                await self.__reqs.put(None)
+                await asyncio.wait_for(self.__reqs.put(None), timeout=60)
                 self.__sent_sentinel.set()
             await asyncio.sleep(refresh_time)
 
diff --git a/weaviate/collections/batch/sync.py b/weaviate/collections/batch/sync.py
index f219de563..eaa96b1c5 100644
--- a/weaviate/collections/batch/sync.py
+++ b/weaviate/collections/batch/sync.py
@@ -206,7 +206,7 @@ def __loop(self) -> None:
                 and not self.__is_shutting_down.is_set()
                 and not self.__is_oom.is_set()
             ):
-                self.__reqs.put(None)
+                self.__reqs.put(None, timeout=60)
                 self.__sent_sentinel.set()
             time.sleep(refresh_time)
 

From fadcebb906644fdba18be771ebbb1d4e4fdeec13 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Fri, 20 Mar 2026 17:00:55 +0000
Subject: [PATCH 25/99] Handle the timeout of sentinel pushing gracefully

---
 weaviate/collections/batch/async_.py | 9 ++++++++-
 weaviate/collections/batch/sync.py   | 9 ++++++++-
 2 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/weaviate/collections/batch/async_.py b/weaviate/collections/batch/async_.py
index b86bc80d3..cb7454b30 100644
--- a/weaviate/collections/batch/async_.py
+++ b/weaviate/collections/batch/async_.py
@@ -254,7 +254,14 @@ async def __loop(self) -> None:
                 and not self.__is_shutting_down.is_set()
                 and not self.__is_oom.is_set()
             ):
-                await asyncio.wait_for(self.__reqs.put(None), timeout=60)
+                try:
+                    await asyncio.wait_for(self.__reqs.put(None), timeout=60)
+                except asyncio.TimeoutError as e:
+                    logger.warning(
+                        "Batch queue is blocked for more than 60 seconds while trying to send shutdown signal. Exiting the loop"
+                    )
+                    self.__bg_exception = e
+                    return
                 self.__sent_sentinel.set()
             await asyncio.sleep(refresh_time)
 
diff --git a/weaviate/collections/batch/sync.py b/weaviate/collections/batch/sync.py
index eaa96b1c5..cf7ac2532 100644
--- a/weaviate/collections/batch/sync.py
+++ b/weaviate/collections/batch/sync.py
@@ -206,7 +206,14 @@ def __loop(self) -> None:
                 and not self.__is_shutting_down.is_set()
                 and not self.__is_oom.is_set()
             ):
-                self.__reqs.put(None, timeout=60)
+                try:
+                    self.__reqs.put(None, timeout=60)
+                except Full as e:
+                    logger.warning(
+                        "Batch queue is blocked for more than 60 seconds while trying to send shutdown signal. Exiting the loop"
+                    )
+                    self.__bg_exception = e
+                    return
                 self.__sent_sentinel.set()
             time.sleep(refresh_time)
 

From 78ff5ecfdd24587f3c1dbc7d8b06b54ce4f164d1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Thu, 9 Apr 2026 14:35:00 +0100
Subject: [PATCH 26/99] feat: add TextAnalyzerConfig for ASCII folding in text
 properties

---
 integration/test_collection_config.py         | 73 +++++++++++++++++
 test/collection/test_config.py                | 80 +++++++++++++++++++
 test/collection/test_config_methods.py        | 80 ++++++++++++++++++-
 weaviate/classes/config.py                    |  2 +
 weaviate/collections/classes/config.py        | 35 ++++++++
 .../collections/classes/config_methods.py     | 13 +++
 6 files changed, 282 insertions(+), 1 deletion(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 371405a1d..4ba57532c 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -30,6 +30,7 @@
     PQEncoderType,
     PQEncoderDistribution,
     StopwordsPreset,
+    TextAnalyzerConfig,
     VectorDistances,
     VectorIndexType,
     Vectorizers,
@@ -2196,3 +2197,75 @@ def test_delete_property_index(
         assert config.properties[0].index_range_filters is False
         assert config.properties[0].index_searchable is _index_searchable
         assert config.properties[0].index_filterable is _index_filterable
+
+
+def test_property_text_analyzer_ascii_fold(collection_factory: CollectionFactory) -> None:
+    """Create a collection with ascii folding configured and verify it round-trips."""
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        properties=[
+            Property(
+                name="title",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(
+                    ascii_fold=True, ascii_fold_ignore=["é"]
+                ),
+            ),
+            Property(
+                name="body",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+            ),
+        ],
+    )
+
+    config = collection.config.get()
+    title = next(p for p in config.properties if p.name == "title")
+    body = next(p for p in config.properties if p.name == "body")
+
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.ascii_fold is True
+    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+
+    # Properties without a text_analyzer should not have one in the parsed config.
+    assert body.text_analyzer is None
+
+    # Folding actually takes effect: 'école' is searchable as 'ecole' but 'é' is preserved.
+    collection.data.insert({"title": "école française", "body": "école française"})
+    res = collection.query.bm25(query="ecole", query_properties=["title"])
+    assert len(res.objects) == 1
+    res = collection.query.bm25(query="ecole", query_properties=["body"])
+    assert len(res.objects) == 0
+
+
+def test_property_text_analyzer_ascii_fold_in_nested_property(
+    collection_factory: CollectionFactory,
+) -> None:
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        properties=[
+            Property(
+                name="meta",
+                data_type=DataType.OBJECT,
+                nested_properties=[
+                    Property(
+                        name="title",
+                        data_type=DataType.TEXT,
+                        tokenization=Tokenization.WORD,
+                        text_analyzer=TextAnalyzerConfig(
+                            ascii_fold=True, ascii_fold_ignore=["ñ"]
+                        ),
+                    ),
+                ],
+            ),
+        ],
+    )
+
+    config = collection.config.get()
+    meta = next(p for p in config.properties if p.name == "meta")
+    assert meta.nested_properties is not None
+    nested_title = next(np for np in meta.nested_properties if np.name == "title")
+    assert nested_title.text_analyzer is not None
+    assert nested_title.text_analyzer.ascii_fold is True
+    assert nested_title.text_analyzer.ascii_fold_ignore == ["ñ"]
diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index 523ddc980..bae46f6c1 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -12,6 +12,8 @@
     Property,
     Reconfigure,
     ReferenceProperty,
+    TextAnalyzerConfig,
+    Tokenization,
     Vectorizers,
     _CollectionConfigCreate,
     _GenerativeProvider,
@@ -3021,3 +3023,81 @@ def test_nested_property_with_id_name_is_allowed() -> None:
         ],
     )
     assert prop.nestedProperties[0].name == "id"
+
+
+class TestTextAnalyzerConfig:
+    def test_property_without_text_analyzer_omits_key(self) -> None:
+        prop = Property(name="title", data_type=DataType.TEXT)
+        assert "textAnalyzer" not in prop._to_dict()
+
+    def test_property_with_ascii_fold_only(self) -> None:
+        prop = Property(
+            name="title",
+            data_type=DataType.TEXT,
+            text_analyzer=TextAnalyzerConfig(ascii_fold=True),
+        )
+        assert prop._to_dict()["textAnalyzer"] == {"asciiFold": True}
+
+    def test_property_with_ascii_fold_and_ignore(self) -> None:
+        prop = Property(
+            name="title",
+            data_type=DataType.TEXT,
+            tokenization=Tokenization.WORD,
+            text_analyzer=TextAnalyzerConfig(
+                ascii_fold=True, ascii_fold_ignore=["é", "ñ"]
+            ),
+        )
+        out = prop._to_dict()
+        assert out["textAnalyzer"] == {
+            "asciiFold": True,
+            "asciiFoldIgnore": ["é", "ñ"],
+        }
+        assert out["tokenization"] == "word"
+
+    def test_text_analyzer_default_omits_unset_fields(self) -> None:
+        prop = Property(
+            name="title",
+            data_type=DataType.TEXT,
+            text_analyzer=TextAnalyzerConfig(),
+        )
+        # exclude_none drops both unset fields, leaving an empty dict
+        assert prop._to_dict()["textAnalyzer"] == {}
+
+    def test_text_analyzer_only_ignore_list(self) -> None:
+        prop = Property(
+            name="title",
+            data_type=DataType.TEXT,
+            text_analyzer=TextAnalyzerConfig(ascii_fold_ignore=["é"]),
+        )
+        assert prop._to_dict()["textAnalyzer"] == {"asciiFoldIgnore": ["é"]}
+
+    def test_nested_property_with_text_analyzer(self) -> None:
+        prop = Property(
+            name="meta",
+            data_type=DataType.OBJECT,
+            nested_properties=[
+                Property(
+                    name="title",
+                    data_type=DataType.TEXT,
+                    text_analyzer=TextAnalyzerConfig(
+                        ascii_fold=True, ascii_fold_ignore=["ñ"]
+                    ),
+                ),
+            ],
+        )
+        out = prop._to_dict()
+        assert out["nestedProperties"][0]["textAnalyzer"] == {
+            "asciiFold": True,
+            "asciiFoldIgnore": ["ñ"],
+        }
+
+    def test_text_analyzer_accepts_snake_case_alias(self) -> None:
+        ta = TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"])
+        assert ta.asciiFold is True
+        assert ta.asciiFoldIgnore == ["é"]
+
+    def test_text_analyzer_rejects_wrong_types(self) -> None:
+        with pytest.raises(ValidationError):
+            TextAnalyzerConfig(ascii_fold="yes")  # type: ignore[arg-type]
+        with pytest.raises(ValidationError):
+            TextAnalyzerConfig(ascii_fold_ignore="é")  # type: ignore[arg-type]
diff --git a/test/collection/test_config_methods.py b/test/collection/test_config_methods.py
index fbc33b702..d934053c9 100644
--- a/test/collection/test_config_methods.py
+++ b/test/collection/test_config_methods.py
@@ -1,4 +1,8 @@
-from weaviate.collections.classes.config_methods import _collection_configs_simple_from_json
+from weaviate.collections.classes.config_methods import (
+    _collection_configs_simple_from_json,
+    _nested_properties_from_config,
+    _properties_from_config,
+)
 
 
 def test_collection_config_simple_from_json_with_none_vectorizer_config() -> None:
@@ -68,3 +72,77 @@ def test_collection_config_simple_from_json_with_none_vectorizer_config() -> Non
     assert "default" in vec_config
     assert vec_config["default"].vectorizer.model == {}
     assert vec_config["default"].vectorizer.source_properties is None
+
+
+def _make_text_prop(name: str, **extra) -> dict:
+    base = {
+        "name": name,
+        "dataType": ["text"],
+        "indexFilterable": True,
+        "indexSearchable": True,
+        "indexRangeFilters": False,
+        "tokenization": "word",
+    }
+    base.update(extra)
+    return base
+
+
+def test_properties_from_config_parses_text_analyzer() -> None:
+    schema = {
+        "vectorizer": "none",
+        "properties": [
+            _make_text_prop(
+                "title",
+                textAnalyzer={"asciiFold": True, "asciiFoldIgnore": ["é"]},
+            ),
+            _make_text_prop("body"),
+        ],
+    }
+    props = _properties_from_config(schema)
+    title = next(p for p in props if p.name == "title")
+    body = next(p for p in props if p.name == "body")
+
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.ascii_fold is True
+    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+
+    assert body.text_analyzer is None
+
+    # The dataclass round-trips back to the wire format.
+    assert title.to_dict()["textAnalyzer"] == {
+        "asciiFold": True,
+        "asciiFoldIgnore": ["é"],
+    }
+    assert "textAnalyzer" not in body.to_dict()
+
+
+def test_properties_from_config_text_analyzer_defaults_when_partial() -> None:
+    schema = {
+        "vectorizer": "none",
+        "properties": [
+            _make_text_prop("title", textAnalyzer={"asciiFoldIgnore": ["é"]}),
+        ],
+    }
+    title = _properties_from_config(schema)[0]
+    assert title.text_analyzer is not None
+    # asciiFold defaults to False when omitted from the server response
+    assert title.text_analyzer.ascii_fold is False
+    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+
+
+def test_nested_properties_from_config_parses_text_analyzer() -> None:
+    nested = _nested_properties_from_config(
+        [
+            _make_text_prop(
+                "title",
+                textAnalyzer={"asciiFold": True, "asciiFoldIgnore": ["ñ"]},
+            ),
+        ]
+    )
+    assert nested[0].text_analyzer is not None
+    assert nested[0].text_analyzer.ascii_fold is True
+    assert nested[0].text_analyzer.ascii_fold_ignore == ["ñ"]
+    assert nested[0].to_dict()["textAnalyzer"] == {
+        "asciiFold": True,
+        "asciiFoldIgnore": ["ñ"],
+    }
diff --git a/weaviate/classes/config.py b/weaviate/classes/config.py
index ce1faf993..868cd1c79 100644
--- a/weaviate/classes/config.py
+++ b/weaviate/classes/config.py
@@ -12,6 +12,7 @@
     ReplicationDeletionStrategy,
     Rerankers,
     StopwordsPreset,
+    TextAnalyzerConfig,
     Tokenization,
     VectorDistances,
 )
@@ -39,6 +40,7 @@
     "ReferenceProperty",
     "Rerankers",
     "StopwordsPreset",
+    "TextAnalyzerConfig",
     "Tokenization",
     "Vectorizers",
     "VectorDistances",
diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 2a8b8d600..415d7cc47 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -1671,6 +1671,12 @@ class _PropertyVectorizerConfig:
 PropertyVectorizerConfig = _PropertyVectorizerConfig
 
 
+@dataclass
+class _TextAnalyzerConfig(_ConfigBase):
+    ascii_fold: bool
+    ascii_fold_ignore: Optional[List[str]]
+
+
 @dataclass
 class _NestedProperty(_ConfigBase):
     data_type: DataType
@@ -1679,6 +1685,7 @@ class _NestedProperty(_ConfigBase):
     index_searchable: bool
     name: str
     nested_properties: Optional[List["NestedProperty"]]
+    text_analyzer: Optional[_TextAnalyzerConfig]
     tokenization: Optional[Tokenization]
 
     def to_dict(self) -> Dict[str, Any]:
@@ -1712,6 +1719,7 @@ class _Property(_PropertyBase):
     index_range_filters: bool
     index_searchable: bool
     nested_properties: Optional[List[NestedProperty]]
+    text_analyzer: Optional[_TextAnalyzerConfig]
     tokenization: Optional[Tokenization]
     vectorizer_config: Optional[PropertyVectorizerConfig]
     vectorizer: Optional[str]
@@ -1724,6 +1732,8 @@ def to_dict(self) -> Dict[str, Any]:
         out["indexSearchable"] = self.index_searchable
         out["indexRangeFilters"] = self.index_range_filters
         out["tokenization"] = self.tokenization.value if self.tokenization else None
+        if self.text_analyzer is not None:
+            out["textAnalyzer"] = self.text_analyzer.to_dict()
         if self.nested_properties is not None and len(self.nested_properties) > 0:
             out["nestedProperties"] = [np.to_dict() for np in self.nested_properties]
         module_config: Dict[str, Any] = {}
@@ -2161,6 +2171,27 @@ class _ShardStatus:
 ShardStatus = _ShardStatus
 
 
+class TextAnalyzerConfig(_ConfigCreateModel):
+    """Text analysis options for a property.
+
+    Configures ASCII folding behavior for `text` and `text[]` properties that use an
+    inverted index (searchable or filterable). When enabled, accent/diacritic marks are
+    folded to their base characters during indexing and search (e.g. 'école' matches
+    'ecole').
+
+    Attributes:
+        ascii_fold: If True, accent/diacritic marks are folded to their base characters
+            during indexing and search. Defaults to False.
+        ascii_fold_ignore: Optional list of characters that should be excluded from
+            ASCII folding (e.g. ['é'] keeps 'é' from being folded to 'e').
+
+    Both settings are immutable after the property is created.
+    """
+
+    asciiFold: Optional[bool] = Field(default=None, alias="ascii_fold")
+    asciiFoldIgnore: Optional[List[str]] = Field(default=None, alias="ascii_fold_ignore")
+
+
 class Property(_ConfigCreateModel):
     """This class defines the structure of a data property that a collection can have within Weaviate.
 
@@ -2173,6 +2204,9 @@ class Property(_ConfigCreateModel):
         index_searchable: Whether the property should be searchable in the inverted index.
         nested_properties: nested properties for data type OBJECT and OBJECT_ARRAY`.
         skip_vectorization: Whether to skip vectorization of the property. Defaults to `False`.
+        text_analyzer: Text analysis options for the property. Configures ASCII folding
+            behavior for text and text[] properties using an inverted index. Immutable
+            after the property is created.
         tokenization: The tokenization method to use for the inverted index. Defaults to `None`.
         vectorize_property_name: Whether to vectorize the property name. Defaults to `True`.
     """
@@ -2187,6 +2221,7 @@ class Property(_ConfigCreateModel):
         default=None, alias="nested_properties"
     )
     skip_vectorization: bool = Field(default=False)
+    textAnalyzer: Optional[TextAnalyzerConfig] = Field(default=None, alias="text_analyzer")
     tokenization: Optional[Tokenization] = Field(default=None)
     vectorize_property_name: bool = Field(default=True)
 
diff --git a/weaviate/collections/classes/config_methods.py b/weaviate/collections/classes/config_methods.py
index c150394f1..4bb8a53b9 100644
--- a/weaviate/collections/classes/config_methods.py
+++ b/weaviate/collections/classes/config_methods.py
@@ -39,6 +39,7 @@
     _ShardingConfig,
     _SQConfig,
     _StopwordsConfig,
+    _TextAnalyzerConfig,
     _VectorIndexConfigDynamic,
     _VectorIndexConfigFlat,
     _VectorIndexConfigHFresh,
@@ -462,6 +463,16 @@ def _collection_configs_simple_from_json(
     return dict(sorted(configs.items()))
 
 
+def _text_analyzer_from_config(prop: Dict[str, Any]) -> Optional[_TextAnalyzerConfig]:
+    ta = prop.get("textAnalyzer")
+    if ta is None:
+        return None
+    return _TextAnalyzerConfig(
+        ascii_fold=ta.get("asciiFold", False),
+        ascii_fold_ignore=ta.get("asciiFoldIgnore"),
+    )
+
+
 def _nested_properties_from_config(props: List[Dict[str, Any]]) -> List[_NestedProperty]:
     return [
         _NestedProperty(
@@ -475,6 +486,7 @@ def _nested_properties_from_config(props: List[Dict[str, Any]]) -> List[_NestedP
                 if prop.get("nestedProperties") is not None
                 else None
             ),
+            text_analyzer=_text_analyzer_from_config(prop),
             tokenization=(
                 Tokenization(prop["tokenization"]) if prop.get("tokenization") is not None else None
             ),
@@ -497,6 +509,7 @@ def _properties_from_config(schema: Dict[str, Any]) -> List[_Property]:
                 if prop.get("nestedProperties") is not None
                 else None
             ),
+            text_analyzer=_text_analyzer_from_config(prop),
             tokenization=(
                 Tokenization(prop["tokenization"]) if prop.get("tokenization") is not None else None
             ),

From 6931a6f180ebb6955cb2ad799b7c84fdd5141b27 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Thu, 9 Apr 2026 14:38:13 +0100
Subject: [PATCH 27/99] refactor: ruff format

---
 integration/test_collection_config.py | 8 ++------
 test/collection/test_config.py        | 8 ++------
 2 files changed, 4 insertions(+), 12 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 4ba57532c..99305adb7 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2208,9 +2208,7 @@ def test_property_text_analyzer_ascii_fold(collection_factory: CollectionFactory
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(
-                    ascii_fold=True, ascii_fold_ignore=["é"]
-                ),
+                text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"]),
             ),
             Property(
                 name="body",
@@ -2253,9 +2251,7 @@ def test_property_text_analyzer_ascii_fold_in_nested_property(
                         name="title",
                         data_type=DataType.TEXT,
                         tokenization=Tokenization.WORD,
-                        text_analyzer=TextAnalyzerConfig(
-                            ascii_fold=True, ascii_fold_ignore=["ñ"]
-                        ),
+                        text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["ñ"]),
                     ),
                 ],
             ),
diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index bae46f6c1..4693a44aa 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -3043,9 +3043,7 @@ def test_property_with_ascii_fold_and_ignore(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=TextAnalyzerConfig(
-                ascii_fold=True, ascii_fold_ignore=["é", "ñ"]
-            ),
+            text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é", "ñ"]),
         )
         out = prop._to_dict()
         assert out["textAnalyzer"] == {
@@ -3079,9 +3077,7 @@ def test_nested_property_with_text_analyzer(self) -> None:
                 Property(
                     name="title",
                     data_type=DataType.TEXT,
-                    text_analyzer=TextAnalyzerConfig(
-                        ascii_fold=True, ascii_fold_ignore=["ñ"]
-                    ),
+                    text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["ñ"]),
                 ),
             ],
         )

From bda3008db236c891d6f67094ca2bdd70ba00ab9a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Thu, 9 Apr 2026 14:46:58 +0100
Subject: [PATCH 28/99] feat: add min version check

---
 integration/test_collection_config.py        | 36 +++++++++++++++++++-
 weaviate/collections/collections/executor.py | 22 +++++++++++-
 weaviate/collections/config/executor.py      | 20 +++++++++++
 3 files changed, 76 insertions(+), 2 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 99305adb7..ef1705b39 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -43,7 +43,11 @@
     IndexName,
 )
 from weaviate.collections.classes.tenants import Tenant
-from weaviate.exceptions import UnexpectedStatusCodeError, WeaviateInvalidInputError
+from weaviate.exceptions import (
+    UnexpectedStatusCodeError,
+    WeaviateInvalidInputError,
+    WeaviateUnsupportedFeatureError,
+)
 from integration.conftest import retry_on_http_error
 
 
@@ -2201,6 +2205,10 @@ def test_delete_property_index(
 
 def test_property_text_analyzer_ascii_fold(collection_factory: CollectionFactory) -> None:
     """Create a collection with ascii folding configured and verify it round-trips."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("Property text_analyzer (asciiFold) requires Weaviate >= 1.37.0")
+
     collection = collection_factory(
         vectorizer_config=Configure.Vectorizer.none(),
         properties=[
@@ -2240,6 +2248,10 @@ def test_property_text_analyzer_ascii_fold(collection_factory: CollectionFactory
 def test_property_text_analyzer_ascii_fold_in_nested_property(
     collection_factory: CollectionFactory,
 ) -> None:
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("Property text_analyzer (asciiFold) requires Weaviate >= 1.37.0")
+
     collection = collection_factory(
         vectorizer_config=Configure.Vectorizer.none(),
         properties=[
@@ -2265,3 +2277,25 @@ def test_property_text_analyzer_ascii_fold_in_nested_property(
     assert nested_title.text_analyzer is not None
     assert nested_title.text_analyzer.ascii_fold is True
     assert nested_title.text_analyzer.ascii_fold_ignore == ["ñ"]
+
+
+def test_property_text_analyzer_ascii_fold_version_gate(
+    collection_factory: CollectionFactory,
+) -> None:
+    """On Weaviate < 1.37 the client must raise before sending the request."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_at_least(1, 37, 0):
+        pytest.skip("Version gate only applies to Weaviate < 1.37.0")
+
+    with pytest.raises(WeaviateUnsupportedFeatureError):
+        collection_factory(
+            vectorizer_config=Configure.Vectorizer.none(),
+            properties=[
+                Property(
+                    name="title",
+                    data_type=DataType.TEXT,
+                    tokenization=Tokenization.WORD,
+                    text_analyzer=TextAnalyzerConfig(ascii_fold=True),
+                ),
+            ],
+        )
diff --git a/weaviate/collections/collections/executor.py b/weaviate/collections/collections/executor.py
index 8497cdf51..eb90df5d7 100644
--- a/weaviate/collections/collections/executor.py
+++ b/weaviate/collections/collections/executor.py
@@ -50,7 +50,7 @@
     ConnectionType,
     _ExpectedStatusCodes,
 )
-from weaviate.exceptions import WeaviateInvalidInputError
+from weaviate.exceptions import WeaviateInvalidInputError, WeaviateUnsupportedFeatureError
 from weaviate.util import _capitalize_first_letter, _decode_json_response_dict
 from weaviate.validator import _validate_input, _ValidateArgument
 from weaviate.warnings import _Warnings
@@ -58,6 +58,19 @@
 CollectionType = TypeVar("CollectionType", Collection, CollectionAsync)
 
 
+def _any_property_has_text_analyzer(properties: Sequence[Property]) -> bool:
+    for prop in properties:
+        if prop.textAnalyzer is not None:
+            return True
+        nested = prop.nestedProperties
+        if nested is None:
+            continue
+        nested_list = nested if isinstance(nested, list) else [nested]
+        if _any_property_has_text_analyzer(nested_list):
+            return True
+    return False
+
+
 class _CollectionsExecutor(Generic[ConnectionType]):
     def __init__(self, connection: ConnectionType):
         self._connection = connection
@@ -213,6 +226,13 @@ def create(
             _Warnings.vectorizer_config_in_config_create()
         if vector_index_config is not None:
             _Warnings.vector_index_config_in_config_create()
+        if properties is not None and _any_property_has_text_analyzer(properties):
+            if not self._connection._weaviate_version.is_at_least(1, 37, 0):
+                raise WeaviateUnsupportedFeatureError(
+                    "Property text_analyzer (asciiFold)",
+                    str(self._connection._weaviate_version),
+                    "1.37.0",
+                )
         try:
             config = _CollectionConfigCreate(
                 description=description,
diff --git a/weaviate/collections/config/executor.py b/weaviate/collections/config/executor.py
index c95cba5a3..92dc1d792 100644
--- a/weaviate/collections/config/executor.py
+++ b/weaviate/collections/config/executor.py
@@ -53,6 +53,7 @@
 from weaviate.connect.v4 import ConnectionAsync, ConnectionType, _ExpectedStatusCodes
 from weaviate.exceptions import (
     WeaviateInvalidInputError,
+    WeaviateUnsupportedFeatureError,
 )
 from weaviate.util import (
     _capitalize_first_letter,
@@ -63,6 +64,16 @@
 from weaviate.warnings import _Warnings
 
 
+def _property_has_text_analyzer(prop: Property) -> bool:
+    if prop.textAnalyzer is not None:
+        return True
+    nested = prop.nestedProperties
+    if nested is None:
+        return False
+    nested_list = nested if isinstance(nested, list) else [nested]
+    return any(_property_has_text_analyzer(np) for np in nested_list)
+
+
 class _ConfigCollectionExecutor(Generic[ConnectionType]):
     def __init__(
         self,
@@ -244,6 +255,15 @@ async def _execute() -> None:
         return executor.result(resp(schema))
 
     def __add_property(self, additional_property: PropertyType) -> executor.Result[None]:
+        if isinstance(additional_property, Property) and _property_has_text_analyzer(
+            additional_property
+        ):
+            if not self._connection._weaviate_version.is_at_least(1, 37, 0):
+                raise WeaviateUnsupportedFeatureError(
+                    "Property text_analyzer (asciiFold)",
+                    str(self._connection._weaviate_version),
+                    "1.37.0",
+                )
         path = f"/schema/{self._name}/properties"
         obj = additional_property._to_dict()
 

From 77fc0ffc9c57d6a83a52a8ffa313c4ff9b3f0079 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Thu, 9 Apr 2026 21:50:42 +0100
Subject: [PATCH 29/99] feat: update TextAnalyzerConfig docstring for
 ascii_fold attributes

---
 weaviate/collections/classes/config.py | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 415d7cc47..e15f33dc6 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -2181,9 +2181,11 @@ class TextAnalyzerConfig(_ConfigCreateModel):
 
     Attributes:
         ascii_fold: If True, accent/diacritic marks are folded to their base characters
-            during indexing and search. Defaults to False.
+            during indexing and search. If omitted, the field is not sent to the server
+            and the server default (False) applies.
         ascii_fold_ignore: Optional list of characters that should be excluded from
-            ASCII folding (e.g. ['é'] keeps 'é' from being folded to 'e').
+            ASCII folding (e.g. ['é'] keeps 'é' from being folded to 'e'). If omitted,
+            the field is not sent to the server.
 
     Both settings are immutable after the property is created.
     """

From a8d6927dd79953dc4800461ab163cd823b5b39a1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Thu, 9 Apr 2026 21:50:51 +0100
Subject: [PATCH 30/99] feat: add asciiFold check in _text_analyzer_from_config
 function

---
 weaviate/collections/classes/config_methods.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/weaviate/collections/classes/config_methods.py b/weaviate/collections/classes/config_methods.py
index 4bb8a53b9..e79ff45e9 100644
--- a/weaviate/collections/classes/config_methods.py
+++ b/weaviate/collections/classes/config_methods.py
@@ -467,6 +467,8 @@ def _text_analyzer_from_config(prop: Dict[str, Any]) -> Optional[_TextAnalyzerCo
     ta = prop.get("textAnalyzer")
     if ta is None:
         return None
+    if "asciiFold" not in ta:
+        return None
     return _TextAnalyzerConfig(
         ascii_fold=ta.get("asciiFold", False),
         ascii_fold_ignore=ta.get("asciiFoldIgnore"),

From e8919a3a9c2778ced02bda5927443aa7dc34aa8f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Thu, 9 Apr 2026 21:51:07 +0100
Subject: [PATCH 31/99] test: fix ASCII folding tests

---
 integration/test_collection_config.py  | 65 ++++++++++++++++++++------
 test/collection/test_config_methods.py |  9 ++--
 2 files changed, 56 insertions(+), 18 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index ef1705b39..6a4cd70b0 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2212,14 +2212,23 @@ def test_property_text_analyzer_ascii_fold(collection_factory: CollectionFactory
     collection = collection_factory(
         vectorizer_config=Configure.Vectorizer.none(),
         properties=[
+            # Folds all accents.
             Property(
-                name="title",
+                name="folded",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(ascii_fold=True),
+            ),
+            # Folds all accents EXCEPT 'é'.
+            Property(
+                name="folded_except_e_acute",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
                 text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"]),
             ),
+            # No folding at all.
             Property(
-                name="body",
+                name="plain",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
             ),
@@ -2227,21 +2236,51 @@ def test_property_text_analyzer_ascii_fold(collection_factory: CollectionFactory
     )
 
     config = collection.config.get()
-    title = next(p for p in config.properties if p.name == "title")
-    body = next(p for p in config.properties if p.name == "body")
+    folded = next(p for p in config.properties if p.name == "folded")
+    folded_ignore = next(p for p in config.properties if p.name == "folded_except_e_acute")
+    plain = next(p for p in config.properties if p.name == "plain")
+
+    assert folded.text_analyzer is not None
+    assert folded.text_analyzer.ascii_fold is True
+    assert folded.text_analyzer.ascii_fold_ignore is None
+
+    assert folded_ignore.text_analyzer is not None
+    assert folded_ignore.text_analyzer.ascii_fold is True
+    assert folded_ignore.text_analyzer.ascii_fold_ignore == ["é"]
+
+    assert plain.text_analyzer is None
 
-    assert title.text_analyzer is not None
-    assert title.text_analyzer.ascii_fold is True
-    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+    # Insert one doc that contains both an ignored accent ('é') and a non-ignored
+    # accent ('ç') so the test can distinguish "folded" from "ignored" behavior.
+    collection.data.insert(
+        {
+            "folded": "école française",
+            "folded_except_e_acute": "école française",
+            "plain": "école française",
+        }
+    )
 
-    # Properties without a text_analyzer should not have one in the parsed config.
-    assert body.text_analyzer is None
+    # `folded` folds everything → both 'ecole' and 'francaise' match.
+    res = collection.query.bm25(query="ecole", query_properties=["folded"])
+    assert len(res.objects) == 1
+    res = collection.query.bm25(query="francaise", query_properties=["folded"])
+    assert len(res.objects) == 1
 
-    # Folding actually takes effect: 'école' is searchable as 'ecole' but 'é' is preserved.
-    collection.data.insert({"title": "école française", "body": "école française"})
-    res = collection.query.bm25(query="ecole", query_properties=["title"])
+    # `folded_except_e_acute` preserves 'é' but still folds 'ç' → 'c'.
+    # Querying 'ecole' must NOT match because 'é' is in the ignore list.
+    res = collection.query.bm25(query="ecole", query_properties=["folded_except_e_acute"])
+    assert len(res.objects) == 0
+    # Querying the original 'école' still matches.
+    res = collection.query.bm25(query="école", query_properties=["folded_except_e_acute"])
     assert len(res.objects) == 1
-    res = collection.query.bm25(query="ecole", query_properties=["body"])
+    # 'ç' is not in the ignore list, so it still folds and 'francaise' matches.
+    res = collection.query.bm25(query="francaise", query_properties=["folded_except_e_acute"])
+    assert len(res.objects) == 1
+
+    # `plain` does no folding at all → ASCII queries don't match accented tokens.
+    res = collection.query.bm25(query="ecole", query_properties=["plain"])
+    assert len(res.objects) == 0
+    res = collection.query.bm25(query="francaise", query_properties=["plain"])
     assert len(res.objects) == 0
 
 
diff --git a/test/collection/test_config_methods.py b/test/collection/test_config_methods.py
index d934053c9..472213ff9 100644
--- a/test/collection/test_config_methods.py
+++ b/test/collection/test_config_methods.py
@@ -116,18 +116,17 @@ def test_properties_from_config_parses_text_analyzer() -> None:
     assert "textAnalyzer" not in body.to_dict()
 
 
-def test_properties_from_config_text_analyzer_defaults_when_partial() -> None:
+def test_properties_from_config_text_analyzer_omitted_when_no_ascii_fold() -> None:
+    """If the server response omits asciiFold, the client treats text_analyzer as unset."""
     schema = {
         "vectorizer": "none",
         "properties": [
+            # Server response with textAnalyzer present but no asciiFold key
             _make_text_prop("title", textAnalyzer={"asciiFoldIgnore": ["é"]}),
         ],
     }
     title = _properties_from_config(schema)[0]
-    assert title.text_analyzer is not None
-    # asciiFold defaults to False when omitted from the server response
-    assert title.text_analyzer.ascii_fold is False
-    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+    assert title.text_analyzer is None
 
 
 def test_nested_properties_from_config_parses_text_analyzer() -> None:

From 3cc63063de362d03a9a8cb5b40c3496ebc6c62ec Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Thu, 9 Apr 2026 22:09:42 +0100
Subject: [PATCH 32/99] feat: add support for stopword presets in inverted
 index configuration and text analyzer

---
 weaviate/collections/classes/config.py        | 56 +++++++++++++++----
 weaviate/collections/classes/config_base.py   |  2 +-
 .../collections/classes/config_methods.py     |  6 +-
 weaviate/collections/collections/executor.py  | 12 +++-
 weaviate/collections/config/executor.py       | 10 ++++
 5 files changed, 73 insertions(+), 13 deletions(-)

diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index e15f33dc6..3d837b81c 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -380,12 +380,14 @@ class _InvertedIndexConfigCreate(_ConfigCreateModel):
     indexPropertyLength: Optional[bool]
     indexNullState: Optional[bool]
     stopwords: _StopwordsCreate
+    stopwordPresets: Optional[Dict[str, List[str]]] = None
 
 
 class _InvertedIndexConfigUpdate(_ConfigUpdateModel):
     bm25: Optional[_BM25ConfigUpdate]
     cleanupIntervalSeconds: Optional[int]
     stopwords: Optional[_StopwordsUpdate]
+    stopwordPresets: Optional[Dict[str, List[str]]] = None
 
 
 class _MultiTenancyConfigCreate(_ConfigCreateModel):
@@ -1647,6 +1649,7 @@ class _InvertedIndexConfig(_ConfigBase):
     index_property_length: bool
     index_timestamps: bool
     stopwords: StopwordsConfig
+    stopword_presets: Optional[Dict[str, List[str]]] = None
 
 
 InvertedIndexConfig = _InvertedIndexConfig
@@ -1675,6 +1678,7 @@ class _PropertyVectorizerConfig:
 class _TextAnalyzerConfig(_ConfigBase):
     ascii_fold: bool
     ascii_fold_ignore: Optional[List[str]]
+    stopword_preset: Optional[str]
 
 
 @dataclass
@@ -2174,24 +2178,42 @@ class _ShardStatus:
 class TextAnalyzerConfig(_ConfigCreateModel):
     """Text analysis options for a property.
 
-    Configures ASCII folding behavior for `text` and `text[]` properties that use an
-    inverted index (searchable or filterable). When enabled, accent/diacritic marks are
-    folded to their base characters during indexing and search (e.g. 'école' matches
-    'ecole').
+    Configures per-property text analysis for `text` and `text[]` properties that use an
+    inverted index (searchable or filterable). Supports ASCII folding (accent/diacritic
+    handling) and selecting a stopword preset that overrides the collection-level
+    `invertedIndexConfig.stopwords` setting for this property only.
 
     Attributes:
         ascii_fold: If True, accent/diacritic marks are folded to their base characters
-            during indexing and search. If omitted, the field is not sent to the server
-            and the server default (False) applies.
+            during indexing and search (e.g. 'école' matches 'ecole'). If omitted, the
+            field is not sent to the server and the server default (False) applies.
         ascii_fold_ignore: Optional list of characters that should be excluded from
             ASCII folding (e.g. ['é'] keeps 'é' from being folded to 'e'). If omitted,
             the field is not sent to the server.
-
-    Both settings are immutable after the property is created.
+        stopword_preset: Stopword preset name. Overrides the collection-level
+            `invertedIndexConfig.stopwords` for this property. Only applies to
+            properties using `Tokenization.WORD`. Accepts a built-in preset
+            (`StopwordsPreset.EN` or `StopwordsPreset.NONE`) or the name of a
+            user-defined preset declared in
+            `Configure.inverted_index(stopword_presets=...)`.
+
+    All settings are immutable after the property is created.
     """
 
     asciiFold: Optional[bool] = Field(default=None, alias="ascii_fold")
     asciiFoldIgnore: Optional[List[str]] = Field(default=None, alias="ascii_fold_ignore")
+    stopwordPreset: Optional[Union[StopwordsPreset, str]] = Field(
+        default=None, alias="stopword_preset"
+    )
+
+    @field_validator("stopwordPreset", mode="before")
+    @classmethod
+    def _coerce_stopword_preset(cls, v: Any) -> Any:
+        # Pydantic preserves the StopwordsPreset enum instance through model_dump,
+        # but the wire format must be a plain string. Coerce at construction time.
+        if isinstance(v, StopwordsPreset):
+            return v.value
+        return v
 
 
 class Property(_ConfigCreateModel):
@@ -2615,11 +2637,17 @@ def inverted_index(
         stopwords_preset: Optional[StopwordsPreset] = None,
         stopwords_additions: Optional[List[str]] = None,
         stopwords_removals: Optional[List[str]] = None,
+        stopword_presets: Optional[Dict[str, List[str]]] = None,
     ) -> _InvertedIndexConfigCreate:
         """Create an `InvertedIndexConfigCreate` object to be used when defining the configuration of the keyword searching algorithm of Weaviate.
 
         Args:
-            See [the docs](https://weaviate.io/developers/weaviate/configuration/indexes#configure-the-inverted-index) for details!
+            stopword_presets: User-defined named stopword lists keyed by preset name. Each value
+                is a flat list of stopword strings. A preset can be referenced from a property's
+                `text_analyzer.stopword_preset` to override the collection-level stopwords for
+                that property only. Requires Weaviate >= 1.37.0.
+
+            See [the docs](https://weaviate.io/developers/weaviate/configuration/indexes#configure-the-inverted-index) for details on the other parameters.
         """  # noqa: D417 (missing argument descriptions in the docstring)
         if bm25_b is None and bm25_k1 is not None or bm25_k1 is None and bm25_b is not None:
             raise ValueError("bm25_b and bm25_k1 must be specified together")
@@ -2639,6 +2667,7 @@ def inverted_index(
                 additions=stopwords_additions,
                 removals=stopwords_removals,
             ),
+            stopwordPresets=stopword_presets,
         )
 
     @staticmethod
@@ -2913,13 +2942,19 @@ def inverted_index(
         stopwords_additions: Optional[List[str]] = None,
         stopwords_preset: Optional[StopwordsPreset] = None,
         stopwords_removals: Optional[List[str]] = None,
+        stopword_presets: Optional[Dict[str, List[str]]] = None,
     ) -> _InvertedIndexConfigUpdate:
         """Create an `InvertedIndexConfigUpdate` object.
 
         Use this method when defining the `inverted_index_config` argument in `collection.update()`.
 
         Args:
-            See [the docs](https://weaviate.io/developers/weaviate/configuration/indexes#configure-the-inverted-index) for a more detailed view!
+            stopword_presets: User-defined named stopword lists keyed by preset name. Each value
+                is a flat list of stopword strings. Passing this replaces the entire user-defined
+                stopword preset map for the collection. Removing a preset still referenced by a
+                property is rejected by the server. Requires Weaviate >= 1.37.0.
+
+            See [the docs](https://weaviate.io/developers/weaviate/configuration/indexes#configure-the-inverted-index) for details on the other parameters.
         """  # noqa: D417 (missing argument descriptions in the docstring)
         return _InvertedIndexConfigUpdate(
             bm25=_BM25ConfigUpdate(b=bm25_b, k1=bm25_k1),
@@ -2929,6 +2964,7 @@ def inverted_index(
                 additions=stopwords_additions,
                 removals=stopwords_removals,
             ),
+            stopwordPresets=stopword_presets,
         )
 
     @staticmethod
diff --git a/weaviate/collections/classes/config_base.py b/weaviate/collections/classes/config_base.py
index fc696fdfb..aa572795e 100644
--- a/weaviate/collections/classes/config_base.py
+++ b/weaviate/collections/classes/config_base.py
@@ -29,7 +29,7 @@ def merge_with_existing(self, schema: Dict[str, Any]) -> Dict[str, Any]:
                 continue
             if isinstance(val, Enum):
                 schema[cls_field] = str(val.value)
-            elif isinstance(val, (int, float, bool, str, list)):
+            elif isinstance(val, (int, float, bool, str, list, dict)):
                 schema[cls_field] = val
             elif isinstance(val, _QuantizerConfigUpdate):
                 quantizers = ["pq", "bq", "sq"]
diff --git a/weaviate/collections/classes/config_methods.py b/weaviate/collections/classes/config_methods.py
index e79ff45e9..691cf208d 100644
--- a/weaviate/collections/classes/config_methods.py
+++ b/weaviate/collections/classes/config_methods.py
@@ -357,6 +357,7 @@ def _collection_config_from_json(schema: Dict[str, Any]) -> _CollectionConfig:
                 additions=schema["invertedIndexConfig"]["stopwords"]["additions"],
                 removals=schema["invertedIndexConfig"]["stopwords"]["removals"],
             ),
+            stopword_presets=schema["invertedIndexConfig"].get("stopwordPresets"),
         ),
         multi_tenancy_config=_MultiTenancyConfig(
             enabled=schema.get("multiTenancyConfig", {}).get("enabled", False),
@@ -467,11 +468,14 @@ def _text_analyzer_from_config(prop: Dict[str, Any]) -> Optional[_TextAnalyzerCo
     ta = prop.get("textAnalyzer")
     if ta is None:
         return None
-    if "asciiFold" not in ta:
+    # The server normalizes an empty TextAnalyzer to nil (see usecases/schema/validation.go),
+    # so the only meaningful signal is the presence of one of the configured fields.
+    if "asciiFold" not in ta and "stopwordPreset" not in ta:
         return None
     return _TextAnalyzerConfig(
         ascii_fold=ta.get("asciiFold", False),
         ascii_fold_ignore=ta.get("asciiFoldIgnore"),
+        stopword_preset=ta.get("stopwordPreset"),
     )
 
 
diff --git a/weaviate/collections/collections/executor.py b/weaviate/collections/collections/executor.py
index eb90df5d7..cdd46fda9 100644
--- a/weaviate/collections/collections/executor.py
+++ b/weaviate/collections/collections/executor.py
@@ -229,10 +229,20 @@ def create(
         if properties is not None and _any_property_has_text_analyzer(properties):
             if not self._connection._weaviate_version.is_at_least(1, 37, 0):
                 raise WeaviateUnsupportedFeatureError(
-                    "Property text_analyzer (asciiFold)",
+                    "Property text_analyzer (asciiFold / stopword_preset)",
                     str(self._connection._weaviate_version),
                     "1.37.0",
                 )
+        if (
+            inverted_index_config is not None
+            and inverted_index_config.stopwordPresets is not None
+            and not self._connection._weaviate_version.is_at_least(1, 37, 0)
+        ):
+            raise WeaviateUnsupportedFeatureError(
+                "InvertedIndexConfig stopword_presets",
+                str(self._connection._weaviate_version),
+                "1.37.0",
+            )
         try:
             config = _CollectionConfigCreate(
                 description=description,
diff --git a/weaviate/collections/config/executor.py b/weaviate/collections/config/executor.py
index 92dc1d792..ca431198e 100644
--- a/weaviate/collections/config/executor.py
+++ b/weaviate/collections/config/executor.py
@@ -210,6 +210,16 @@ def update(
             ),
         ):
             _Warnings.vectorizer_config_in_config_update()
+        if (
+            inverted_index_config is not None
+            and inverted_index_config.stopwordPresets is not None
+            and not self._connection._weaviate_version.is_at_least(1, 37, 0)
+        ):
+            raise WeaviateUnsupportedFeatureError(
+                "InvertedIndexConfig stopword_presets",
+                str(self._connection._weaviate_version),
+                "1.37.0",
+            )
         try:
             config = _CollectionConfigUpdate(
                 description=description,

From ef04dea6478be2c17635a750038b340219eb5b34 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Thu, 9 Apr 2026 22:10:04 +0100
Subject: [PATCH 33/99] test: added live and config tests

---
 integration/test_collection_config.py  | 569 +++++++++++++++++++++++++
 test/collection/test_config.py         |  98 +++++
 test/collection/test_config_methods.py | 100 +++++
 3 files changed, 767 insertions(+)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 6a4cd70b0..f53667b3e 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -42,6 +42,7 @@
     _VectorizerConfigCreate,
     IndexName,
 )
+from weaviate.collections.classes.filters import Filter
 from weaviate.collections.classes.tenants import Tenant
 from weaviate.exceptions import (
     UnexpectedStatusCodeError,
@@ -2338,3 +2339,571 @@ def test_property_text_analyzer_ascii_fold_version_gate(
                 ),
             ],
         )
+
+
+def test_collection_stopword_presets(collection_factory: CollectionFactory) -> None:
+    """User-defined stopword presets defined at collection level apply to properties
+    that reference them via text_analyzer.stopword_preset, and built-in presets can
+    coexist with user-defined ones."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        inverted_index_config=Configure.inverted_index(
+            stopword_presets={"fr": ["le", "la", "les"]},
+        ),
+        properties=[
+            # User-defined French preset.
+            Property(
+                name="title_fr",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+            ),
+            # Built-in English preset, set per property.
+            Property(
+                name="title_en",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(stopword_preset=StopwordsPreset.EN),
+            ),
+            # No stopword override → uses the collection-level default.
+            Property(
+                name="plain",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+            ),
+        ],
+    )
+
+    config = collection.config.get()
+    assert config.inverted_index_config.stopword_presets == {"fr": ["le", "la", "les"]}
+
+    title_fr = next(p for p in config.properties if p.name == "title_fr")
+    title_en = next(p for p in config.properties if p.name == "title_en")
+    plain = next(p for p in config.properties if p.name == "plain")
+    assert title_fr.text_analyzer is not None
+    assert title_fr.text_analyzer.stopword_preset == "fr"
+    assert title_en.text_analyzer is not None
+    assert title_en.text_analyzer.stopword_preset == "en"
+    assert plain.text_analyzer is None
+
+    collection.data.insert(
+        {
+            "title_fr": "le chat noir",
+            "title_en": "the black cat",
+            "plain": "the black cat",
+        }
+    )
+
+    # title_fr filters 'le' (in user preset) → BM25 for 'le' yields no match,
+    # but 'chat' still matches.
+    res = collection.query.bm25(query="le", query_properties=["title_fr"])
+    assert len(res.objects) == 0
+    res = collection.query.bm25(query="chat", query_properties=["title_fr"])
+    assert len(res.objects) == 1
+
+    # title_en filters 'the' (built-in en) but matches 'cat'.
+    res = collection.query.bm25(query="the", query_properties=["title_en"])
+    assert len(res.objects) == 0
+    res = collection.query.bm25(query="cat", query_properties=["title_en"])
+    assert len(res.objects) == 1
+
+
+def test_collection_stopword_presets_update(collection_factory: CollectionFactory) -> None:
+    """Updating the contents of an existing user-defined stopword preset takes effect
+    immediately for queries against properties that reference it."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        inverted_index_config=Configure.inverted_index(
+            stopword_presets={"fr": ["le"]},
+        ),
+        properties=[
+            Property(
+                name="title_fr",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+            ),
+        ],
+    )
+
+    collection.data.insert({"title_fr": "le chat et la souris"})
+
+    # Baseline: 'le' is filtered, 'la' is not (it isn't in the preset yet).
+    res = collection.query.bm25(query="le", query_properties=["title_fr"])
+    assert len(res.objects) == 0
+    res = collection.query.bm25(query="la", query_properties=["title_fr"])
+    assert len(res.objects) == 1
+
+    # Replace the preset contents: drop 'le', add 'la'.
+    collection.config.update(
+        inverted_index_config=Reconfigure.inverted_index(
+            stopword_presets={"fr": ["la"]},
+        ),
+    )
+
+    config = collection.config.get()
+    assert config.inverted_index_config.stopword_presets == {"fr": ["la"]}
+
+    # After update: 'le' now passes through, 'la' is filtered.
+    res = collection.query.bm25(query="le", query_properties=["title_fr"])
+    assert len(res.objects) == 1
+    res = collection.query.bm25(query="la", query_properties=["title_fr"])
+    assert len(res.objects) == 0
+
+
+def test_collection_stopword_presets_remove_in_use_is_rejected(
+    collection_factory: CollectionFactory,
+) -> None:
+    """The server rejects removing a stopword preset still referenced by a property."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        inverted_index_config=Configure.inverted_index(
+            stopword_presets={"fr": ["le", "la", "les"]},
+        ),
+        properties=[
+            Property(
+                name="title_fr",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+            ),
+        ],
+    )
+
+    with pytest.raises(UnexpectedStatusCodeError):
+        collection.config.update(
+            inverted_index_config=Reconfigure.inverted_index(stopword_presets={}),
+        )
+
+    # The original preset must still be present after the rejected update.
+    config = collection.config.get()
+    assert config.inverted_index_config.stopword_presets == {"fr": ["le", "la", "les"]}
+
+
+def test_inverted_index_stopword_presets_version_gate(
+    collection_factory: CollectionFactory,
+) -> None:
+    """On Weaviate < 1.37 the client must raise before sending the request."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_at_least(1, 37, 0):
+        pytest.skip("Version gate only applies to Weaviate < 1.37.0")
+
+    with pytest.raises(WeaviateUnsupportedFeatureError):
+        collection_factory(
+            vectorizer_config=Configure.Vectorizer.none(),
+            inverted_index_config=Configure.inverted_index(
+                stopword_presets={"fr": ["le", "la"]},
+            ),
+        )
+
+
+def test_collection_stopword_presets_remove_unused_is_allowed(
+    collection_factory: CollectionFactory,
+) -> None:
+    """Removing a preset that no property references must succeed."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        inverted_index_config=Configure.inverted_index(
+            stopword_presets={
+                "fr": ["le", "la", "les"],
+                "es": ["el", "la", "los"],
+            },
+        ),
+        properties=[
+            Property(
+                name="title",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+            ),
+        ],
+    )
+
+    # Drop only 'es' (unused). 'fr' is still referenced by title.
+    collection.config.update(
+        inverted_index_config=Reconfigure.inverted_index(
+            stopword_presets={"fr": ["le", "la", "les"]},
+        ),
+    )
+    config = collection.config.get()
+    assert config.inverted_index_config.stopword_presets == {"fr": ["le", "la", "les"]}
+
+
+def test_collection_stopword_presets_remove_referenced_by_nested_property_is_rejected(
+    collection_factory: CollectionFactory,
+) -> None:
+    """A removed preset still referenced by a nested property must be rejected by the server."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        inverted_index_config=Configure.inverted_index(
+            stopword_presets={"fr": ["le", "la", "les"]},
+        ),
+        properties=[
+            Property(
+                name="doc",
+                data_type=DataType.OBJECT,
+                nested_properties=[
+                    Property(
+                        name="body",
+                        data_type=DataType.TEXT,
+                        tokenization=Tokenization.WORD,
+                        text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+                    ),
+                ],
+            ),
+        ],
+    )
+
+    with pytest.raises(UnexpectedStatusCodeError):
+        collection.config.update(
+            inverted_index_config=Reconfigure.inverted_index(stopword_presets={}),
+        )
+
+    # The original preset must still be present after the rejected update.
+    config = collection.config.get()
+    assert config.inverted_index_config.stopword_presets == {"fr": ["le", "la", "les"]}
+
+
+def test_collection_user_defined_stopword_preset_overrides_builtin(
+    collection_factory: CollectionFactory,
+) -> None:
+    """A user-defined preset named 'en' replaces the built-in 'en' for properties of this collection."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        inverted_index_config=Configure.inverted_index(
+            # Shadow the built-in 'en' with a user-defined preset that filters
+            # 'hello' but NOT 'the'.
+            stopword_presets={"en": ["hello"]},
+        ),
+        properties=[
+            Property(
+                name="title",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(stopword_preset="en"),
+            ),
+        ],
+    )
+
+    collection.data.insert({"title": "the quick hello world"})
+
+    # 'the' is no longer filtered (built-in en is overridden) → matches.
+    res = collection.query.bm25(query="the", query_properties=["title"])
+    assert len(res.objects) == 1
+    # 'hello' was added by the override → filtered.
+    res = collection.query.bm25(query="hello", query_properties=["title"])
+    assert len(res.objects) == 0
+    # 'quick' is in neither list → matches.
+    res = collection.query.bm25(query="quick", query_properties=["title"])
+    assert len(res.objects) == 1
+
+
+def test_property_text_analyzer_combined_ascii_fold_and_stopword_preset(
+    collection_factory: CollectionFactory,
+) -> None:
+    """A single property may combine ascii_fold and stopword_preset."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        properties=[
+            Property(
+                name="title",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(
+                    ascii_fold=True,
+                    stopword_preset=StopwordsPreset.EN,
+                ),
+            ),
+        ],
+    )
+
+    config = collection.config.get()
+    title = next(p for p in config.properties if p.name == "title")
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.ascii_fold is True
+    assert title.text_analyzer.stopword_preset == "en"
+
+    collection.data.insert({"title": "The école française"})
+
+    # 'the' is filtered by built-in en → no match.
+    res = collection.query.bm25(query="the", query_properties=["title"])
+    assert len(res.objects) == 0
+    # 'ecole' matches because ascii_fold folds 'école'.
+    res = collection.query.bm25(query="ecole", query_properties=["title"])
+    assert len(res.objects) == 1
+    # 'francaise' matches because ascii_fold folds 'française'.
+    res = collection.query.bm25(query="francaise", query_properties=["title"])
+    assert len(res.objects) == 1
+
+
+def test_property_text_analyzer_ascii_fold_immutable(
+    collection_factory: CollectionFactory,
+) -> None:
+    """The asciiFold setting is immutable; updating it via add_property on a renamed
+    property is the only way to change behavior, but the original property cannot
+    have its analyzer mutated."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        properties=[
+            Property(
+                name="title",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"]),
+            ),
+        ],
+    )
+
+    # The config exposes the original ignore list and there's no client API
+    # surface to mutate text_analyzer on an existing property — it can only be
+    # set at create time.
+    config = collection.config.get()
+    title = next(p for p in config.properties if p.name == "title")
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+
+    # Adding a *new* property with a different analyzer is allowed.
+    collection.config.add_property(
+        Property(
+            name="title2",
+            data_type=DataType.TEXT,
+            tokenization=Tokenization.WORD,
+            text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["ñ"]),
+        ),
+    )
+    config = collection.config.get()
+    title = next(p for p in config.properties if p.name == "title")
+    title2 = next(p for p in config.properties if p.name == "title2")
+    # Original property's analyzer is unchanged.
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+    # New property has its own analyzer.
+    assert title2.text_analyzer is not None
+    assert title2.text_analyzer.ascii_fold_ignore == ["ñ"]
+
+
+@pytest.mark.parametrize(
+    "tokenization,query_match,query_no_match",
+    [
+        (Tokenization.WORD, "école", "ecole"),
+        (Tokenization.LOWERCASE, "l'école", "l'ecole"),
+        (Tokenization.WHITESPACE, "L'école", "L'ecole"),
+        (Tokenization.FIELD, "L'école est fermée", "L'ecole est fermee"),
+        (Tokenization.TRIGRAM, "éco", "eco"),
+    ],
+    ids=["word", "lowercase", "whitespace", "field", "trigram"],
+)
+def test_property_ascii_fold_across_tokenizations(
+    collection_factory: CollectionFactory,
+    tokenization: Tokenization,
+    query_match: str,
+    query_no_match: str,
+) -> None:
+    """ascii_fold + ignore list ['é'] preserves é under every supported tokenization."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        properties=[
+            Property(
+                name="prop",
+                data_type=DataType.TEXT,
+                tokenization=tokenization,
+                text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"]),
+            ),
+            Property(
+                name="prop_no_ignore",
+                data_type=DataType.TEXT,
+                tokenization=tokenization,
+                text_analyzer=TextAnalyzerConfig(ascii_fold=True),
+            ),
+        ],
+    )
+    collection.data.insert(
+        {
+            "prop": "L'école est fermée",
+            "prop_no_ignore": "L'école est fermée",
+        }
+    )
+
+    # On the property with the ignore list, the accented form matches but the
+    # ASCII-folded form does not.
+    res = collection.query.bm25(query=query_match, query_properties=["prop"])
+    assert len(res.objects) == 1
+    res = collection.query.bm25(query=query_no_match, query_properties=["prop"])
+    assert len(res.objects) == 0
+
+    # On the property without the ignore list, both forms match.
+    res = collection.query.bm25(query=query_match, query_properties=["prop_no_ignore"])
+    assert len(res.objects) == 1
+    res = collection.query.bm25(query=query_no_match, query_properties=["prop_no_ignore"])
+    assert len(res.objects) == 1
+
+
+def test_property_ascii_fold_multi_char_ignore(
+    collection_factory: CollectionFactory,
+) -> None:
+    """An ignore list with multiple characters preserves all of them while still
+    folding the rest."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        properties=[
+            Property(
+                name="multi_ignore",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(
+                    ascii_fold=True,
+                    ascii_fold_ignore=["é", "ü", "ñ", "ø"],
+                ),
+            ),
+        ],
+    )
+    collection.data.insert_many(
+        [
+            {"multi_ignore": "résumé"},
+            {"multi_ignore": "über"},
+            {"multi_ignore": "El Niño"},
+            {"multi_ignore": "Ørsted"},
+            {"multi_ignore": "São Paulo"},
+            {"multi_ignore": "naïve café"},
+        ]
+    )
+
+    # All four ignored chars are preserved → ASCII-folded form does not match,
+    # but the accented form does.
+    for accented, folded in [
+        ("résumé", "resume"),
+        ("über", "uber"),
+        ("niño", "nino"),
+        ("ørsted", "orsted"),
+    ]:
+        match = collection.query.bm25(query=accented, query_properties=["multi_ignore"])
+        no_match = collection.query.bm25(query=folded, query_properties=["multi_ignore"])
+        assert len(match.objects) == 1, f"{accented} should match"
+        assert len(no_match.objects) == 0, f"{folded} should not match (char in ignore list)"
+
+    # Non-ignored accents (ã, ï) still fold normally.
+    res = collection.query.bm25(query="sao", query_properties=["multi_ignore"])
+    assert len(res.objects) == 1
+    res = collection.query.bm25(query="naive", query_properties=["multi_ignore"])
+    assert len(res.objects) == 1
+
+    # The "naïve café" doc has both an ignored char (é) and a non-ignored one (ï):
+    # 'naive' matches (ï folded), 'cafe' does not (é preserved).
+    res = collection.query.fetch_objects(
+        filters=Filter.by_property("multi_ignore").equal("naive café"),
+        limit=10,
+    )
+    assert len(res.objects) == 1
+
+
+def test_property_ascii_fold_with_filters(
+    collection_factory: CollectionFactory,
+) -> None:
+    """Equal and Like filters respect ascii_fold and ascii_fold_ignore."""
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        properties=[
+            # Folds everything.
+            Property(
+                name="body",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(ascii_fold=True),
+            ),
+            # Folds everything except 'é'.
+            Property(
+                name="title",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"]),
+            ),
+        ],
+    )
+    collection.data.insert_many(
+        [
+            {"title": "L'école est fermée", "body": "L'école est fermée"},
+            {"title": "cafe résumé", "body": "cafe résumé"},
+        ]
+    )
+
+    # Equal on body (full fold): 'ecole' matches.
+    res = collection.query.fetch_objects(
+        filters=Filter.by_property("body").equal("ecole"),
+        limit=5,
+    )
+    assert len(res.objects) == 1
+
+    # Equal on title (é preserved): 'ecole' does not, but 'école' does.
+    res = collection.query.fetch_objects(
+        filters=Filter.by_property("title").equal("école"),
+        limit=5,
+    )
+    assert len(res.objects) == 1
+    res = collection.query.fetch_objects(
+        filters=Filter.by_property("title").equal("ecole"),
+        limit=5,
+    )
+    assert len(res.objects) == 0
+
+    # Like on body (full fold): 'ecol*' matches.
+    res = collection.query.fetch_objects(
+        filters=Filter.by_property("body").like("ecol*"),
+        limit=5,
+    )
+    assert len(res.objects) == 1
+
+    # Like on title (é preserved): 'écol*' matches, 'ecol*' does not.
+    res = collection.query.fetch_objects(
+        filters=Filter.by_property("title").like("écol*"),
+        limit=5,
+    )
+    assert len(res.objects) == 1
+    res = collection.query.fetch_objects(
+        filters=Filter.by_property("title").like("ecol*"),
+        limit=5,
+    )
+    assert len(res.objects) == 0
diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index 4693a44aa..fcae98d74 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -12,6 +12,7 @@
     Property,
     Reconfigure,
     ReferenceProperty,
+    StopwordsPreset,
     TextAnalyzerConfig,
     Tokenization,
     Vectorizers,
@@ -3097,3 +3098,100 @@ def test_text_analyzer_rejects_wrong_types(self) -> None:
             TextAnalyzerConfig(ascii_fold="yes")  # type: ignore[arg-type]
         with pytest.raises(ValidationError):
             TextAnalyzerConfig(ascii_fold_ignore="é")  # type: ignore[arg-type]
+
+    def test_text_analyzer_stopword_preset_builtin_enum(self) -> None:
+        prop = Property(
+            name="title",
+            data_type=DataType.TEXT,
+            tokenization=Tokenization.WORD,
+            text_analyzer=TextAnalyzerConfig(stopword_preset=StopwordsPreset.EN),
+        )
+        assert prop._to_dict()["textAnalyzer"] == {"stopwordPreset": "en"}
+
+    def test_text_analyzer_stopword_preset_user_defined_string(self) -> None:
+        prop = Property(
+            name="title_fr",
+            data_type=DataType.TEXT,
+            tokenization=Tokenization.WORD,
+            text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+        )
+        assert prop._to_dict()["textAnalyzer"] == {"stopwordPreset": "fr"}
+
+    def test_text_analyzer_combined_ascii_fold_and_stopword_preset(self) -> None:
+        prop = Property(
+            name="title",
+            data_type=DataType.TEXT,
+            tokenization=Tokenization.WORD,
+            text_analyzer=TextAnalyzerConfig(
+                ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset="fr"
+            ),
+        )
+        assert prop._to_dict()["textAnalyzer"] == {
+            "asciiFold": True,
+            "asciiFoldIgnore": ["é"],
+            "stopwordPreset": "fr",
+        }
+
+    def test_text_analyzer_stopword_preset_only_omits_other_keys(self) -> None:
+        prop = Property(
+            name="title",
+            data_type=DataType.TEXT,
+            tokenization=Tokenization.WORD,
+            text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+        )
+        out = prop._to_dict()
+        assert "asciiFold" not in out["textAnalyzer"]
+        assert "asciiFoldIgnore" not in out["textAnalyzer"]
+
+
+class TestInvertedIndexStopwordPresets:
+    def test_configure_inverted_index_with_stopword_presets(self) -> None:
+        ic = Configure.inverted_index(
+            stopword_presets={
+                "fr": ["le", "la", "les"],
+                "es": ["el", "la", "los"],
+            },
+        )
+        out = ic._to_dict()
+        assert out["stopwordPresets"] == {
+            "fr": ["le", "la", "les"],
+            "es": ["el", "la", "los"],
+        }
+
+    def test_configure_inverted_index_without_stopword_presets_omits_key(self) -> None:
+        ic = Configure.inverted_index()
+        assert "stopwordPresets" not in ic._to_dict()
+
+    def test_reconfigure_inverted_index_merges_stopword_presets(self) -> None:
+        rc = Reconfigure.inverted_index(stopword_presets={"fr": ["le", "la"]})
+        existing = {
+            "stopwords": {"preset": "en", "additions": None, "removals": None},
+            "bm25": {"b": 0.75, "k1": 1.2},
+            "cleanupIntervalSeconds": 60,
+        }
+        merged = rc.merge_with_existing(existing)
+        assert merged["stopwordPresets"] == {"fr": ["le", "la"]}
+        # other fields untouched
+        assert merged["stopwords"]["preset"] == "en"
+        assert merged["bm25"]["b"] == 0.75
+
+    def test_reconfigure_inverted_index_replaces_existing_stopword_presets(self) -> None:
+        rc = Reconfigure.inverted_index(stopword_presets={"fr": ["le"]})
+        existing = {
+            "stopwords": {"preset": "en", "additions": None, "removals": None},
+            "stopwordPresets": {"fr": ["le", "la", "les"], "es": ["el"]},
+        }
+        merged = rc.merge_with_existing(existing)
+        # The new value fully replaces the prior dict (this matches the server-side
+        # PUT semantics — see test_tokenize.py::test_remove_unused_preset_is_allowed).
+        assert merged["stopwordPresets"] == {"fr": ["le"]}
+
+    def test_reconfigure_inverted_index_without_stopword_presets_leaves_existing(self) -> None:
+        rc = Reconfigure.inverted_index(bm25_b=0.7, bm25_k1=1.1)
+        existing = {
+            "stopwords": {"preset": "en", "additions": None, "removals": None},
+            "bm25": {"b": 0.75, "k1": 1.2},
+            "stopwordPresets": {"fr": ["le", "la"]},
+        }
+        merged = rc.merge_with_existing(existing)
+        assert merged["stopwordPresets"] == {"fr": ["le", "la"]}
diff --git a/test/collection/test_config_methods.py b/test/collection/test_config_methods.py
index 472213ff9..2e40acacc 100644
--- a/test/collection/test_config_methods.py
+++ b/test/collection/test_config_methods.py
@@ -1,4 +1,5 @@
 from weaviate.collections.classes.config_methods import (
+    _collection_config_from_json,
     _collection_configs_simple_from_json,
     _nested_properties_from_config,
     _properties_from_config,
@@ -145,3 +146,102 @@ def test_nested_properties_from_config_parses_text_analyzer() -> None:
         "asciiFold": True,
         "asciiFoldIgnore": ["ñ"],
     }
+
+
+def test_properties_from_config_parses_stopword_preset_only() -> None:
+    """A property with only stopwordPreset (no asciiFold) must still produce a text_analyzer."""
+    schema = {
+        "vectorizer": "none",
+        "properties": [
+            _make_text_prop("title", textAnalyzer={"stopwordPreset": "fr"}),
+        ],
+    }
+    title = _properties_from_config(schema)[0]
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.ascii_fold is False
+    assert title.text_analyzer.ascii_fold_ignore is None
+    assert title.text_analyzer.stopword_preset == "fr"
+
+
+def test_properties_from_config_parses_combined_text_analyzer() -> None:
+    schema = {
+        "vectorizer": "none",
+        "properties": [
+            _make_text_prop(
+                "title",
+                textAnalyzer={
+                    "asciiFold": True,
+                    "asciiFoldIgnore": ["é"],
+                    "stopwordPreset": "fr",
+                },
+            ),
+        ],
+    }
+    title = _properties_from_config(schema)[0]
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.ascii_fold is True
+    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+    assert title.text_analyzer.stopword_preset == "fr"
+
+
+def _full_schema(class_name: str, **inverted_overrides) -> dict:
+    inverted = {
+        "bm25": {"b": 0.75, "k1": 1.2},
+        "cleanupIntervalSeconds": 60,
+        "stopwords": {"preset": "en", "additions": None, "removals": None},
+    }
+    inverted.update(inverted_overrides)
+    return {
+        "class": class_name,
+        "vectorizer": "none",
+        "properties": [],
+        "invertedIndexConfig": inverted,
+        "replicationConfig": {"factor": 1, "deletionStrategy": "NoAutomatedResolution"},
+        "shardingConfig": {
+            "virtualPerPhysical": 128,
+            "desiredCount": 1,
+            "actualCount": 1,
+            "desiredVirtualCount": 128,
+            "actualVirtualCount": 128,
+            "key": "_id",
+            "strategy": "hash",
+            "function": "murmur3",
+        },
+        "vectorIndexType": "hnsw",
+        "vectorIndexConfig": {
+            "skip": False,
+            "cleanupIntervalSeconds": 300,
+            "maxConnections": 64,
+            "efConstruction": 128,
+            "ef": -1,
+            "dynamicEfMin": 100,
+            "dynamicEfMax": 500,
+            "dynamicEfFactor": 8,
+            "vectorCacheMaxObjects": 1000000000000,
+            "flatSearchCutoff": 40000,
+            "distance": "cosine",
+        },
+    }
+
+
+def test_collection_config_parses_stopword_presets() -> None:
+    """The inverted index config exposes stopwordPresets when present in the schema."""
+    schema = _full_schema(
+        "TestStopwordPresets",
+        stopwordPresets={
+            "fr": ["le", "la", "les"],
+            "es": ["el", "la", "los"],
+        },
+    )
+    full = _collection_config_from_json(schema)
+    assert full.inverted_index_config.stopword_presets == {
+        "fr": ["le", "la", "les"],
+        "es": ["el", "la", "los"],
+    }
+
+
+def test_collection_config_stopword_presets_absent() -> None:
+    """If the server response omits stopwordPresets, the parsed value is None."""
+    schema = _full_schema("TestNoStopwordPresets")
+    full = _collection_config_from_json(schema)
+    assert full.inverted_index_config.stopword_presets is None

From 8f1b33b8fe5868860258f5516cf92f7420ae34c2 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Thu, 9 Apr 2026 22:13:19 +0100
Subject: [PATCH 34/99] refactor: improve docstrings for stopword presets and
 asciiFold tests

---
 integration/test_collection_config.py | 29 ++++++++++++++++++---------
 1 file changed, 19 insertions(+), 10 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index f53667b3e..b5ff89b93 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2342,9 +2342,11 @@ def test_property_text_analyzer_ascii_fold_version_gate(
 
 
 def test_collection_stopword_presets(collection_factory: CollectionFactory) -> None:
-    """User-defined stopword presets defined at collection level apply to properties
-    that reference them via text_analyzer.stopword_preset, and built-in presets can
-    coexist with user-defined ones."""
+    """User-defined stopword presets apply to properties that reference them.
+
+    Properties can reference user-defined presets via text_analyzer.stopword_preset,
+    and built-in presets can coexist with user-defined ones.
+    """
     dummy = collection_factory("dummy")
     if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
         pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
@@ -2413,8 +2415,11 @@ def test_collection_stopword_presets(collection_factory: CollectionFactory) -> N
 
 
 def test_collection_stopword_presets_update(collection_factory: CollectionFactory) -> None:
-    """Updating the contents of an existing user-defined stopword preset takes effect
-    immediately for queries against properties that reference it."""
+    """Updating the contents of an existing stopword preset takes effect immediately.
+
+    Queries against properties that reference the preset see the new contents on the
+    next request.
+    """
     dummy = collection_factory("dummy")
     if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
         pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
@@ -2667,9 +2672,11 @@ def test_property_text_analyzer_combined_ascii_fold_and_stopword_preset(
 def test_property_text_analyzer_ascii_fold_immutable(
     collection_factory: CollectionFactory,
 ) -> None:
-    """The asciiFold setting is immutable; updating it via add_property on a renamed
-    property is the only way to change behavior, but the original property cannot
-    have its analyzer mutated."""
+    """The asciiFold setting is immutable on an existing property.
+
+    Adding a new property via add_property is the only way to introduce a different
+    analyzer; the original property's analyzer cannot be mutated.
+    """
     dummy = collection_factory("dummy")
     if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
         pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
@@ -2777,8 +2784,10 @@ def test_property_ascii_fold_across_tokenizations(
 def test_property_ascii_fold_multi_char_ignore(
     collection_factory: CollectionFactory,
 ) -> None:
-    """An ignore list with multiple characters preserves all of them while still
-    folding the rest."""
+    """An ignore list with multiple characters preserves all of them.
+
+    Characters not in the ignore list are still folded normally.
+    """
     dummy = collection_factory("dummy")
     if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
         pytest.skip("text_analyzer requires Weaviate >= 1.37.0")

From 03d6ff44172d81111b6cfc1601f867d36540fd46 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 13 Apr 2026 14:14:09 +0100
Subject: [PATCH 35/99] refactor: simplify _any_property_has_text_analyzer
 function using _property_has_text_analyzer

---
 weaviate/collections/collections/executor.py | 14 +-------------
 weaviate/collections/config/executor.py      |  5 +++++
 2 files changed, 6 insertions(+), 13 deletions(-)

diff --git a/weaviate/collections/collections/executor.py b/weaviate/collections/collections/executor.py
index eb90df5d7..6b73cb53a 100644
--- a/weaviate/collections/collections/executor.py
+++ b/weaviate/collections/collections/executor.py
@@ -50,6 +50,7 @@
     ConnectionType,
     _ExpectedStatusCodes,
 )
+from weaviate.collections.config.executor import _any_property_has_text_analyzer
 from weaviate.exceptions import WeaviateInvalidInputError, WeaviateUnsupportedFeatureError
 from weaviate.util import _capitalize_first_letter, _decode_json_response_dict
 from weaviate.validator import _validate_input, _ValidateArgument
@@ -58,19 +59,6 @@
 CollectionType = TypeVar("CollectionType", Collection, CollectionAsync)
 
 
-def _any_property_has_text_analyzer(properties: Sequence[Property]) -> bool:
-    for prop in properties:
-        if prop.textAnalyzer is not None:
-            return True
-        nested = prop.nestedProperties
-        if nested is None:
-            continue
-        nested_list = nested if isinstance(nested, list) else [nested]
-        if _any_property_has_text_analyzer(nested_list):
-            return True
-    return False
-
-
 class _CollectionsExecutor(Generic[ConnectionType]):
     def __init__(self, connection: ConnectionType):
         self._connection = connection
diff --git a/weaviate/collections/config/executor.py b/weaviate/collections/config/executor.py
index 92dc1d792..51adbd8a0 100644
--- a/weaviate/collections/config/executor.py
+++ b/weaviate/collections/config/executor.py
@@ -5,6 +5,7 @@
     Generic,
     List,
     Literal,
+    Sequence,
     Optional,
     Tuple,
     Union,
@@ -64,6 +65,10 @@
 from weaviate.warnings import _Warnings
 
 
+def _any_property_has_text_analyzer(properties: Sequence[Property]) -> bool:
+    return any(_property_has_text_analyzer(p) for p in properties)
+
+
 def _property_has_text_analyzer(prop: Property) -> bool:
     if prop.textAnalyzer is not None:
         return True

From 1342204274cc29829e115660d279b2dbd987cba1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 13 Apr 2026 14:26:27 +0100
Subject: [PATCH 36/99] test: remove redundant insertion ascii fold tests from
 test_collection_config

---
 integration/test_collection_config.py | 115 --------------------------
 1 file changed, 115 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 6a4cd70b0..93ac923a0 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2203,121 +2203,6 @@ def test_delete_property_index(
         assert config.properties[0].index_filterable is _index_filterable
 
 
-def test_property_text_analyzer_ascii_fold(collection_factory: CollectionFactory) -> None:
-    """Create a collection with ascii folding configured and verify it round-trips."""
-    dummy = collection_factory("dummy")
-    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
-        pytest.skip("Property text_analyzer (asciiFold) requires Weaviate >= 1.37.0")
-
-    collection = collection_factory(
-        vectorizer_config=Configure.Vectorizer.none(),
-        properties=[
-            # Folds all accents.
-            Property(
-                name="folded",
-                data_type=DataType.TEXT,
-                tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(ascii_fold=True),
-            ),
-            # Folds all accents EXCEPT 'é'.
-            Property(
-                name="folded_except_e_acute",
-                data_type=DataType.TEXT,
-                tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"]),
-            ),
-            # No folding at all.
-            Property(
-                name="plain",
-                data_type=DataType.TEXT,
-                tokenization=Tokenization.WORD,
-            ),
-        ],
-    )
-
-    config = collection.config.get()
-    folded = next(p for p in config.properties if p.name == "folded")
-    folded_ignore = next(p for p in config.properties if p.name == "folded_except_e_acute")
-    plain = next(p for p in config.properties if p.name == "plain")
-
-    assert folded.text_analyzer is not None
-    assert folded.text_analyzer.ascii_fold is True
-    assert folded.text_analyzer.ascii_fold_ignore is None
-
-    assert folded_ignore.text_analyzer is not None
-    assert folded_ignore.text_analyzer.ascii_fold is True
-    assert folded_ignore.text_analyzer.ascii_fold_ignore == ["é"]
-
-    assert plain.text_analyzer is None
-
-    # Insert one doc that contains both an ignored accent ('é') and a non-ignored
-    # accent ('ç') so the test can distinguish "folded" from "ignored" behavior.
-    collection.data.insert(
-        {
-            "folded": "école française",
-            "folded_except_e_acute": "école française",
-            "plain": "école française",
-        }
-    )
-
-    # `folded` folds everything → both 'ecole' and 'francaise' match.
-    res = collection.query.bm25(query="ecole", query_properties=["folded"])
-    assert len(res.objects) == 1
-    res = collection.query.bm25(query="francaise", query_properties=["folded"])
-    assert len(res.objects) == 1
-
-    # `folded_except_e_acute` preserves 'é' but still folds 'ç' → 'c'.
-    # Querying 'ecole' must NOT match because 'é' is in the ignore list.
-    res = collection.query.bm25(query="ecole", query_properties=["folded_except_e_acute"])
-    assert len(res.objects) == 0
-    # Querying the original 'école' still matches.
-    res = collection.query.bm25(query="école", query_properties=["folded_except_e_acute"])
-    assert len(res.objects) == 1
-    # 'ç' is not in the ignore list, so it still folds and 'francaise' matches.
-    res = collection.query.bm25(query="francaise", query_properties=["folded_except_e_acute"])
-    assert len(res.objects) == 1
-
-    # `plain` does no folding at all → ASCII queries don't match accented tokens.
-    res = collection.query.bm25(query="ecole", query_properties=["plain"])
-    assert len(res.objects) == 0
-    res = collection.query.bm25(query="francaise", query_properties=["plain"])
-    assert len(res.objects) == 0
-
-
-def test_property_text_analyzer_ascii_fold_in_nested_property(
-    collection_factory: CollectionFactory,
-) -> None:
-    dummy = collection_factory("dummy")
-    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
-        pytest.skip("Property text_analyzer (asciiFold) requires Weaviate >= 1.37.0")
-
-    collection = collection_factory(
-        vectorizer_config=Configure.Vectorizer.none(),
-        properties=[
-            Property(
-                name="meta",
-                data_type=DataType.OBJECT,
-                nested_properties=[
-                    Property(
-                        name="title",
-                        data_type=DataType.TEXT,
-                        tokenization=Tokenization.WORD,
-                        text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["ñ"]),
-                    ),
-                ],
-            ),
-        ],
-    )
-
-    config = collection.config.get()
-    meta = next(p for p in config.properties if p.name == "meta")
-    assert meta.nested_properties is not None
-    nested_title = next(np for np in meta.nested_properties if np.name == "title")
-    assert nested_title.text_analyzer is not None
-    assert nested_title.text_analyzer.ascii_fold is True
-    assert nested_title.text_analyzer.ascii_fold_ignore == ["ñ"]
-
-
 def test_property_text_analyzer_ascii_fold_version_gate(
     collection_factory: CollectionFactory,
 ) -> None:

From cb53d6a0e630b1f32694063272c0c36934a2ef1f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 13 Apr 2026 14:31:02 +0100
Subject: [PATCH 37/99] test: add stopwords roundtrip test for collection
 configuration

---
 integration/test_collection_config.py | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 93ac923a0..43b3f6ee1 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2223,3 +2223,26 @@ def test_property_text_analyzer_ascii_fold_version_gate(
                 ),
             ],
         )
+
+
+def test_stopwords_roundtrip_from_dict(collection_factory: CollectionFactory) -> None:
+    collection = collection_factory(
+        inverted_index_config=Configure.inverted_index(
+            stopwords_additions=["a"],
+            stopwords_preset=StopwordsPreset.EN,
+            stopwords_removals=["the"],
+        ),
+    )
+    config = collection.config.get()
+    assert config.inverted_index_config.stopwords.preset == StopwordsPreset.EN
+    assert config.inverted_index_config.stopwords.removals == ["the"]
+
+    name = f"TestStopwordsRoundtrip{collection.name}"
+    config.name = name
+    with weaviate.connect_to_local() as client:
+        client.collections.delete(name)
+        client.collections.create_from_dict(config.to_dict())
+        new = client.collections.use(name).config.get()
+        assert config == new
+        assert config.to_dict() == new.to_dict()
+        client.collections.delete(name)

From 9de03f3d62f75e5f8ed02fc2f7939904bbc049ce Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 13 Apr 2026 14:32:25 +0100
Subject: [PATCH 38/99] feat: add model validator to enforce asciiFoldIgnore
 constraints in TextAnalyzerConfig

---
 weaviate/collections/classes/config.py | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index e15f33dc6..33d0d7858 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -15,7 +15,7 @@
 )
 
 from deprecation import deprecated as docstring_deprecated
-from pydantic import AnyHttpUrl, Field, TypeAdapter, ValidationInfo, field_validator
+from pydantic import AnyHttpUrl, Field, TypeAdapter, ValidationInfo, field_validator, model_validator
 from typing_extensions import TypeAlias
 from typing_extensions import deprecated as typing_deprecated
 
@@ -2193,6 +2193,12 @@ class TextAnalyzerConfig(_ConfigCreateModel):
     asciiFold: Optional[bool] = Field(default=None, alias="ascii_fold")
     asciiFoldIgnore: Optional[List[str]] = Field(default=None, alias="ascii_fold_ignore")
 
+    @model_validator(mode="after")
+    def _validate_ascii_fold_ignore(self) -> "TextAnalyzerConfig":
+        if self.asciiFold is not True and self.asciiFoldIgnore is not None:
+            raise ValueError("asciiFoldIgnore cannot be set when asciiFold is not enabled")
+        return self
+
 
 class Property(_ConfigCreateModel):
     """This class defines the structure of a data property that a collection can have within Weaviate.

From 7018927e01cc177ce10fc1d1c360d93cd4f1a8c3 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 13 Apr 2026 14:39:55 +0100
Subject: [PATCH 39/99] feat: add factory class for text analyzer
 configurations with ASCII folding support

---
 weaviate/collections/classes/config.py | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 33d0d7858..677258860 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -2200,6 +2200,25 @@ def _validate_ascii_fold_ignore(self) -> "TextAnalyzerConfig":
         return self
 
 
+class _TextAnalyzer:
+    """Factory class for creating text analyzer configurations.
+
+    Use ``Configure.TextAnalyzer`` to access these methods.
+    """
+
+    @staticmethod
+    def ascii_fold(
+        ignore: Optional[List[str]] = None,
+    ) -> _TextAnalyzerConfig:
+        """Create a text analyzer config with ASCII folding enabled.
+
+        Args:
+            ignore: Optional list of characters that should be excluded from
+                ASCII folding (e.g. ``['é']`` keeps 'é' from being folded to 'e').
+        """
+        return _TextAnalyzerConfig(asciiFold=True, asciiFoldIgnore=ignore)
+
+
 class Property(_ConfigCreateModel):
     """This class defines the structure of a data property that a collection can have within Weaviate.
 
@@ -2609,6 +2628,7 @@ class Configure:
     MultiVectors = _MultiVectors
     ObjectTTL = _ObjectTTL
     Replication = _Replication
+    TextAnalyzer = _TextAnalyzer
 
     @staticmethod
     def inverted_index(

From 8e919846e3703abbed7ea8249e8f16cb46deb74e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 13 Apr 2026 17:44:03 +0100
Subject: [PATCH 40/99] refactor: update TextAnalyzerConfig usage to new
 Configure class methods

---
 integration/test_collection_config.py  |  3 +-
 test/collection/test_config.py         | 44 ++++++++------------------
 weaviate/collections/classes/config.py | 22 +++++++++----
 3 files changed, 30 insertions(+), 39 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 43b3f6ee1..ddd059b45 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -30,7 +30,6 @@
     PQEncoderType,
     PQEncoderDistribution,
     StopwordsPreset,
-    TextAnalyzerConfig,
     VectorDistances,
     VectorIndexType,
     Vectorizers,
@@ -2219,7 +2218,7 @@ def test_property_text_analyzer_ascii_fold_version_gate(
                     name="title",
                     data_type=DataType.TEXT,
                     tokenization=Tokenization.WORD,
-                    text_analyzer=TextAnalyzerConfig(ascii_fold=True),
+                    text_analyzer=Configure.TextAnalyzer.ascii_fold(),
                 ),
             ],
         )
diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index 4693a44aa..4acf6bcd3 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -4,20 +4,20 @@
 from pydantic import ValidationError
 
 from weaviate.collections.classes.config import (
-    _AsyncReplicationConfig,
-    _ReplicationConfig,
-    _ReplicationConfigUpdate,
     Configure,
     DataType,
     Property,
     Reconfigure,
     ReferenceProperty,
-    TextAnalyzerConfig,
     Tokenization,
     Vectorizers,
+    _AsyncReplicationConfig,
     _CollectionConfigCreate,
     _GenerativeProvider,
+    _ReplicationConfig,
+    _ReplicationConfigUpdate,
     _RerankerProvider,
+    _TextAnalyzerConfigCreate,
     _VectorizerConfigCreate,
     _ReplicationConfigCreate,
     ReplicationDeletionStrategy,
@@ -3025,7 +3025,7 @@ def test_nested_property_with_id_name_is_allowed() -> None:
     assert prop.nestedProperties[0].name == "id"
 
 
-class TestTextAnalyzerConfig:
+class Test_TextAnalyzerConfigCreate:
     def test_property_without_text_analyzer_omits_key(self) -> None:
         prop = Property(name="title", data_type=DataType.TEXT)
         assert "textAnalyzer" not in prop._to_dict()
@@ -3034,7 +3034,7 @@ def test_property_with_ascii_fold_only(self) -> None:
         prop = Property(
             name="title",
             data_type=DataType.TEXT,
-            text_analyzer=TextAnalyzerConfig(ascii_fold=True),
+            text_analyzer=Configure.TextAnalyzer.ascii_fold(),
         )
         assert prop._to_dict()["textAnalyzer"] == {"asciiFold": True}
 
@@ -3043,7 +3043,7 @@ def test_property_with_ascii_fold_and_ignore(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é", "ñ"]),
+            text_analyzer=Configure.TextAnalyzer.ascii_fold(ignore=["é", "ñ"]),
         )
         out = prop._to_dict()
         assert out["textAnalyzer"] == {
@@ -3052,22 +3052,9 @@ def test_property_with_ascii_fold_and_ignore(self) -> None:
         }
         assert out["tokenization"] == "word"
 
-    def test_text_analyzer_default_omits_unset_fields(self) -> None:
-        prop = Property(
-            name="title",
-            data_type=DataType.TEXT,
-            text_analyzer=TextAnalyzerConfig(),
-        )
-        # exclude_none drops both unset fields, leaving an empty dict
-        assert prop._to_dict()["textAnalyzer"] == {}
-
-    def test_text_analyzer_only_ignore_list(self) -> None:
-        prop = Property(
-            name="title",
-            data_type=DataType.TEXT,
-            text_analyzer=TextAnalyzerConfig(ascii_fold_ignore=["é"]),
-        )
-        assert prop._to_dict()["textAnalyzer"] == {"asciiFoldIgnore": ["é"]}
+    def test_text_analyzer_rejects_ignore_without_ascii_fold(self) -> None:
+        with pytest.raises(ValidationError):
+            _TextAnalyzerConfigCreate(ascii_fold_ignore=["é"])
 
     def test_nested_property_with_text_analyzer(self) -> None:
         prop = Property(
@@ -3077,7 +3064,7 @@ def test_nested_property_with_text_analyzer(self) -> None:
                 Property(
                     name="title",
                     data_type=DataType.TEXT,
-                    text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["ñ"]),
+                    text_analyzer=Configure.TextAnalyzer.ascii_fold(ignore=["ñ"]),
                 ),
             ],
         )
@@ -3087,13 +3074,8 @@ def test_nested_property_with_text_analyzer(self) -> None:
             "asciiFoldIgnore": ["ñ"],
         }
 
-    def test_text_analyzer_accepts_snake_case_alias(self) -> None:
-        ta = TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"])
-        assert ta.asciiFold is True
-        assert ta.asciiFoldIgnore == ["é"]
-
     def test_text_analyzer_rejects_wrong_types(self) -> None:
         with pytest.raises(ValidationError):
-            TextAnalyzerConfig(ascii_fold="yes")  # type: ignore[arg-type]
+            _TextAnalyzerConfigCreate(ascii_fold="yes")  # type: ignore[arg-type]
         with pytest.raises(ValidationError):
-            TextAnalyzerConfig(ascii_fold_ignore="é")  # type: ignore[arg-type]
+            _TextAnalyzerConfigCreate(ascii_fold_ignore="é")  # type: ignore[arg-type]
diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 677258860..6d2dd42cd 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -15,7 +15,14 @@
 )
 
 from deprecation import deprecated as docstring_deprecated
-from pydantic import AnyHttpUrl, Field, TypeAdapter, ValidationInfo, field_validator, model_validator
+from pydantic import (
+    AnyHttpUrl,
+    Field,
+    TypeAdapter,
+    ValidationInfo,
+    field_validator,
+    model_validator,
+)
 from typing_extensions import TypeAlias
 from typing_extensions import deprecated as typing_deprecated
 
@@ -1677,6 +1684,9 @@ class _TextAnalyzerConfig(_ConfigBase):
     ascii_fold_ignore: Optional[List[str]]
 
 
+TextAnalyzerConfig = _TextAnalyzerConfig
+
+
 @dataclass
 class _NestedProperty(_ConfigBase):
     data_type: DataType
@@ -2171,7 +2181,7 @@ class _ShardStatus:
 ShardStatus = _ShardStatus
 
 
-class TextAnalyzerConfig(_ConfigCreateModel):
+class _TextAnalyzerConfigCreate(_ConfigCreateModel):
     """Text analysis options for a property.
 
     Configures ASCII folding behavior for `text` and `text[]` properties that use an
@@ -2194,7 +2204,7 @@ class TextAnalyzerConfig(_ConfigCreateModel):
     asciiFoldIgnore: Optional[List[str]] = Field(default=None, alias="ascii_fold_ignore")
 
     @model_validator(mode="after")
-    def _validate_ascii_fold_ignore(self) -> "TextAnalyzerConfig":
+    def _validate_ascii_fold_ignore(self) -> "_TextAnalyzerConfigCreate":
         if self.asciiFold is not True and self.asciiFoldIgnore is not None:
             raise ValueError("asciiFoldIgnore cannot be set when asciiFold is not enabled")
         return self
@@ -2209,14 +2219,14 @@ class _TextAnalyzer:
     @staticmethod
     def ascii_fold(
         ignore: Optional[List[str]] = None,
-    ) -> _TextAnalyzerConfig:
+    ) -> _TextAnalyzerConfigCreate:
         """Create a text analyzer config with ASCII folding enabled.
 
         Args:
             ignore: Optional list of characters that should be excluded from
                 ASCII folding (e.g. ``['é']`` keeps 'é' from being folded to 'e').
         """
-        return _TextAnalyzerConfig(asciiFold=True, asciiFoldIgnore=ignore)
+        return _TextAnalyzerConfigCreate(ascii_fold=True, ascii_fold_ignore=ignore)
 
 
 class Property(_ConfigCreateModel):
@@ -2248,7 +2258,7 @@ class Property(_ConfigCreateModel):
         default=None, alias="nested_properties"
     )
     skip_vectorization: bool = Field(default=False)
-    textAnalyzer: Optional[TextAnalyzerConfig] = Field(default=None, alias="text_analyzer")
+    textAnalyzer: Optional[_TextAnalyzerConfigCreate] = Field(default=None, alias="text_analyzer")
     tokenization: Optional[Tokenization] = Field(default=None)
     vectorize_property_name: bool = Field(default=True)
 

From db3009c935b8da9c62071bc175771d52577c89d4 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 13 Apr 2026 22:20:45 +0100
Subject: [PATCH 41/99] test: remove redundant line in stopword presets merge
 test

---
 test/collection/test_config.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index 3d271693a..6d2bb0e74 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -3178,4 +3178,3 @@ def test_reconfigure_inverted_index_without_stopword_presets_leaves_existing(sel
         }
         merged = rc.merge_with_existing(existing)
         assert merged["stopwordPresets"] == {"fr": ["le", "la"]}
-

From 50f7768224ca402b5dd340f0f7e6969df0788748 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 13 Apr 2026 22:29:05 +0100
Subject: [PATCH 42/99] refactor: use factory pattern

---
 integration/test_collection_config.py  | 32 +++++++++++++-------------
 test/collection/test_config.py         | 15 ++++++------
 weaviate/collections/classes/config.py | 25 +++++++++++++++-----
 3 files changed, 42 insertions(+), 30 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index fd8208ca2..a125a3ef2 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2219,7 +2219,7 @@ def test_property_text_analyzer_ascii_fold_version_gate(
                     name="title",
                     data_type=DataType.TEXT,
                     tokenization=Tokenization.WORD,
-                    text_analyzer=Configure.TextAnalyzer.ascii_fold(),
+                    text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True),
                 ),
             ],
         )
@@ -2246,14 +2246,14 @@ def test_collection_stopword_presets(collection_factory: CollectionFactory) -> N
                 name="title_fr",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
             ),
             # Built-in English preset, set per property.
             Property(
                 name="title_en",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(stopword_preset=StopwordsPreset.EN),
+                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset=StopwordsPreset.EN),
             ),
             # No stopword override → uses the collection-level default.
             Property(
@@ -2318,7 +2318,7 @@ def test_collection_stopword_presets_update(collection_factory: CollectionFactor
                 name="title_fr",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
             ),
         ],
     )
@@ -2366,7 +2366,7 @@ def test_collection_stopword_presets_remove_in_use_is_rejected(
                 name="title_fr",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
             ),
         ],
     )
@@ -2419,7 +2419,7 @@ def test_collection_stopword_presets_remove_unused_is_allowed(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
             ),
         ],
     )
@@ -2456,7 +2456,7 @@ def test_collection_stopword_presets_remove_referenced_by_nested_property_is_rej
                         name="body",
                         data_type=DataType.TEXT,
                         tokenization=Tokenization.WORD,
-                        text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+                        text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
                     ),
                 ],
             ),
@@ -2493,7 +2493,7 @@ def test_collection_user_defined_stopword_preset_overrides_builtin(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(stopword_preset="en"),
+                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="en"),
             ),
         ],
     )
@@ -2526,7 +2526,7 @@ def test_property_text_analyzer_combined_ascii_fold_and_stopword_preset(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(
+                text_analyzer=Configure.TextAnalyzer.custom(
                     ascii_fold=True,
                     stopword_preset=StopwordsPreset.EN,
                 ),
@@ -2572,7 +2572,7 @@ def test_property_text_analyzer_ascii_fold_immutable(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"]),
+                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["é"]),
             ),
         ],
     )
@@ -2591,7 +2591,7 @@ def test_property_text_analyzer_ascii_fold_immutable(
             name="title2",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["ñ"]),
+            text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["ñ"]),
         ),
     )
     config = collection.config.get()
@@ -2634,13 +2634,13 @@ def test_property_ascii_fold_across_tokenizations(
                 name="prop",
                 data_type=DataType.TEXT,
                 tokenization=tokenization,
-                text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"]),
+                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["é"]),
             ),
             Property(
                 name="prop_no_ignore",
                 data_type=DataType.TEXT,
                 tokenization=tokenization,
-                text_analyzer=TextAnalyzerConfig(ascii_fold=True),
+                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True),
             ),
         ],
     )
@@ -2683,7 +2683,7 @@ def test_property_ascii_fold_multi_char_ignore(
                 name="multi_ignore",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(
+                text_analyzer=Configure.TextAnalyzer.custom(
                     ascii_fold=True,
                     ascii_fold_ignore=["é", "ü", "ñ", "ø"],
                 ),
@@ -2745,14 +2745,14 @@ def test_property_ascii_fold_with_filters(
                 name="body",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(ascii_fold=True),
+                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True),
             ),
             # Folds everything except 'é'.
             Property(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=TextAnalyzerConfig(ascii_fold=True, ascii_fold_ignore=["é"]),
+                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["é"]),
             ),
         ],
     )
diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index 6d2bb0e74..67e4f6da3 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -10,7 +10,6 @@
     Reconfigure,
     ReferenceProperty,
     StopwordsPreset,
-    TextAnalyzerConfig,
     Tokenization,
     Vectorizers,
     _AsyncReplicationConfig,
@@ -3036,7 +3035,7 @@ def test_property_with_ascii_fold_only(self) -> None:
         prop = Property(
             name="title",
             data_type=DataType.TEXT,
-            text_analyzer=Configure.TextAnalyzer.ascii_fold(),
+            text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True),
         )
         assert prop._to_dict()["textAnalyzer"] == {"asciiFold": True}
 
@@ -3045,7 +3044,7 @@ def test_property_with_ascii_fold_and_ignore(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer.ascii_fold(ignore=["é", "ñ"]),
+            text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["é", "ñ"]),
         )
         out = prop._to_dict()
         assert out["textAnalyzer"] == {
@@ -3066,7 +3065,7 @@ def test_nested_property_with_text_analyzer(self) -> None:
                 Property(
                     name="title",
                     data_type=DataType.TEXT,
-                    text_analyzer=Configure.TextAnalyzer.ascii_fold(ignore=["ñ"]),
+                    text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["ñ"]),
                 ),
             ],
         )
@@ -3087,7 +3086,7 @@ def test_text_analyzer_stopword_preset_builtin_enum(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=TextAnalyzerConfig(stopword_preset=StopwordsPreset.EN),
+            text_analyzer=Configure.TextAnalyzer.custom(stopword_preset=StopwordsPreset.EN),
         )
         assert prop._to_dict()["textAnalyzer"] == {"stopwordPreset": "en"}
 
@@ -3096,7 +3095,7 @@ def test_text_analyzer_stopword_preset_user_defined_string(self) -> None:
             name="title_fr",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+            text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
         )
         assert prop._to_dict()["textAnalyzer"] == {"stopwordPreset": "fr"}
 
@@ -3105,7 +3104,7 @@ def test_text_analyzer_combined_ascii_fold_and_stopword_preset(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=TextAnalyzerConfig(
+            text_analyzer=Configure.TextAnalyzer.custom(
                 ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset="fr"
             ),
         )
@@ -3120,7 +3119,7 @@ def test_text_analyzer_stopword_preset_only_omits_other_keys(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=TextAnalyzerConfig(stopword_preset="fr"),
+            text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
         )
         out = prop._to_dict()
         assert "asciiFold" not in out["textAnalyzer"]
diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 7658f40cc..3e6867985 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -2239,16 +2239,29 @@ class _TextAnalyzer:
     """
 
     @staticmethod
-    def ascii_fold(
-        ignore: Optional[List[str]] = None,
+    def custom(
+        ascii_fold: Optional[bool] = None,
+        ascii_fold_ignore: Optional[List[str]] = None,
+        stopword_preset: Optional[Union[StopwordsPreset, str]] = None,
     ) -> _TextAnalyzerConfigCreate:
-        """Create a text analyzer config with ASCII folding enabled.
+        """Create a text analyzer config with custom settings.
 
         Args:
-            ignore: Optional list of characters that should be excluded from
-                ASCII folding (e.g. ``['é']`` keeps 'é' from being folded to 'e').
+            ascii_fold: If True, accent/diacritic marks are folded to their base
+                characters during indexing and search (e.g. 'école' matches 'ecole').
+            ascii_fold_ignore: Optional list of characters that should be excluded
+                from ASCII folding (e.g. ``['é']`` keeps 'é' from being folded to
+                'e'). Requires ``ascii_fold=True``.
+            stopword_preset: Stopword preset name to override the collection-level
+                stopwords for this property. Accepts a ``StopwordsPreset`` or a
+                user-defined preset name.
         """
-        return _TextAnalyzerConfigCreate(ascii_fold=True, ascii_fold_ignore=ignore)
+        return _TextAnalyzerConfigCreate(
+            ascii_fold=ascii_fold,
+            ascii_fold_ignore=ascii_fold_ignore,
+            stopword_preset=stopword_preset,
+        )
+
 
 
 class Property(_ConfigCreateModel):

From 6a1b0bc88b5daa97b219a2eabcd3a7dc3c3ec35b Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Tue, 14 Apr 2026 09:06:26 +0200
Subject: [PATCH 43/99] Add MCP permission

---
 integration/test_rbac.py | 38 ++++++++++++++++++++++++++++++++++++++
 weaviate/rbac/models.py  | 38 ++++++++++++++++++++++++++++++++++++++
 2 files changed, 76 insertions(+)

diff --git a/integration/test_rbac.py b/integration/test_rbac.py
index d98d238a7..53206abac 100644
--- a/integration/test_rbac.py
+++ b/integration/test_rbac.py
@@ -14,6 +14,7 @@
     CollectionsPermissionOutput,
     DataPermissionOutput,
     GroupsPermissionOutput,
+    MCPPermissionOutput,
     NodesPermissionOutput,
     Role,
     ReplicatePermissionOutput,
@@ -44,6 +45,7 @@
                 backups_permissions=[
                     BackupsPermissionOutput(collection="Test", actions={Actions.Backups.MANAGE})
                 ],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -62,6 +64,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -84,6 +87,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -104,6 +108,7 @@
                     DataPermissionOutput(collection="*", tenant="*", actions={Actions.Data.CREATE})
                 ],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -137,6 +142,7 @@
                     ),
                 ],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -155,6 +161,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[
                     NodesPermissionOutput(
                         verbosity="verbose", actions={Actions.Nodes.READ}, collection="Test"
@@ -177,6 +184,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[
                     NodesPermissionOutput(
                         verbosity="minimal", actions={Actions.Nodes.READ}, collection="*"
@@ -203,6 +211,7 @@
                 ],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -221,6 +230,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[
                     TenantsPermissionOutput(
@@ -247,6 +257,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[
                     TenantsPermissionOutput(
@@ -290,6 +301,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -310,6 +322,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[
@@ -355,6 +368,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -379,6 +393,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -403,6 +418,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
@@ -410,6 +426,27 @@
             ),
             32,  # Minimum version for alias permissions
         ),
+        (
+            Permissions.mcp(manage=True),
+            Role(
+                name="ManageMCP",
+                alias_permissions=[],
+                cluster_permissions=[],
+                users_permissions=[],
+                collections_permissions=[],
+                roles_permissions=[],
+                data_permissions=[],
+                backups_permissions=[],
+                mcp_permissions=[
+                    MCPPermissionOutput(actions={Actions.MCP.MANAGE})
+                ],
+                nodes_permissions=[],
+                tenants_permissions=[],
+                replicate_permissions=[],
+                groups_permissions=[],
+            ),
+            37,  # Minimum version for MCP permissions
+        ),
         (
             Permissions.Groups.oidc(group="MyGroup", read=True),
             Role(
@@ -421,6 +458,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
+                mcp_permissions=[],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
diff --git a/weaviate/rbac/models.py b/weaviate/rbac/models.py
index df5a230a5..dfdbc48f4 100644
--- a/weaviate/rbac/models.py
+++ b/weaviate/rbac/models.py
@@ -252,6 +252,14 @@ def values() -> List[str]:
         return [action.value for action in BackupsAction]
 
 
+class MCPAction(str, _Action, Enum):
+    MANAGE = "manage_mcp"
+
+    @staticmethod
+    def values() -> List[str]:
+        return [action.value for action in MCPAction]
+
+
 class ReplicateAction(str, _Action, Enum):
     CREATE = "create_replicate"
     READ = "read_replicate"
@@ -407,6 +415,16 @@ def _to_weaviate(self) -> List[WeaviatePermission]:
         ]
 
 
+class _MCPPermission(_Permission[MCPAction]):
+    def _to_weaviate(self) -> List[WeaviatePermission]:
+        return [
+            {
+                "action": action,
+            }
+            for action in self.actions
+        ]
+
+
 class _ClusterPermission(_Permission[ClusterAction]):
     def _to_weaviate(self) -> List[WeaviatePermission]:
         return [
@@ -470,6 +488,10 @@ class BackupsPermissionOutput(_BackupsPermission):
     pass
 
 
+class MCPPermissionOutput(_MCPPermission):
+    pass
+
+
 class NodesPermissionOutput(_NodesPermission):
     pass
 
@@ -486,6 +508,7 @@ class TenantsPermissionOutput(_TenantsPermission):
     RolesPermissionOutput,
     UsersPermissionOutput,
     BackupsPermissionOutput,
+    MCPPermissionOutput,
     NodesPermissionOutput,
     TenantsPermissionOutput,
     ReplicatePermissionOutput,
@@ -507,6 +530,7 @@ class Role(RoleBase):
     roles_permissions: List[RolesPermissionOutput]
     users_permissions: List[UsersPermissionOutput]
     backups_permissions: List[BackupsPermissionOutput]
+    mcp_permissions: List[MCPPermissionOutput]
     nodes_permissions: List[NodesPermissionOutput]
     tenants_permissions: List[TenantsPermissionOutput]
     replicate_permissions: List[ReplicatePermissionOutput]
@@ -522,6 +546,7 @@ def permissions(self) -> List[PermissionsOutputType]:
         permissions.extend(self.roles_permissions)
         permissions.extend(self.users_permissions)
         permissions.extend(self.backups_permissions)
+        permissions.extend(self.mcp_permissions)
         permissions.extend(self.nodes_permissions)
         permissions.extend(self.tenants_permissions)
         permissions.extend(self.replicate_permissions)
@@ -537,6 +562,7 @@ def _from_weaviate_role(cls, role: WeaviateRole) -> "Role":
         roles_permissions: List[RolesPermissionOutput] = []
         data_permissions: List[DataPermissionOutput] = []
         backups_permissions: List[BackupsPermissionOutput] = []
+        mcp_permissions: List[MCPPermissionOutput] = []
         nodes_permissions: List[NodesPermissionOutput] = []
         tenants_permissions: List[TenantsPermissionOutput] = []
         replicate_permissions: List[ReplicatePermissionOutput] = []
@@ -605,6 +631,10 @@ def _from_weaviate_role(cls, role: WeaviateRole) -> "Role":
                             actions={BackupsAction(permission["action"])},
                         )
                     )
+            elif permission["action"] in MCPAction.values():
+                mcp_permissions.append(
+                    MCPPermissionOutput(actions={MCPAction(permission["action"])})
+                )
             elif permission["action"] in NodesAction.values():
                 nodes = permission.get("nodes")
                 if nodes is not None:
@@ -658,6 +688,7 @@ def _from_weaviate_role(cls, role: WeaviateRole) -> "Role":
             groups_permissions=_join_permissions(groups_permissions),
             data_permissions=_join_permissions(data_permissions),
             backups_permissions=_join_permissions(backups_permissions),
+            mcp_permissions=_join_permissions(mcp_permissions),
             nodes_permissions=_join_permissions(nodes_permissions),
             tenants_permissions=_join_permissions(tenants_permissions),
             replicate_permissions=_join_permissions(replicate_permissions),
@@ -710,6 +741,7 @@ class Actions:
     Cluster = ClusterAction
     Nodes = NodesAction
     Backups = BackupsAction
+    MCP = MCPAction
     Tenants = TenantsAction
     Users = UsersAction
     Replicate = ReplicateAction
@@ -1020,6 +1052,12 @@ def backup(
                 permissions.append(permission)
         return permissions
 
+    @staticmethod
+    def mcp(*, manage: bool = False) -> PermissionsCreateType:
+        if manage:
+            return [_MCPPermission(actions={MCPAction.MANAGE})]
+        return []
+
     @staticmethod
     def cluster(*, read: bool = False) -> PermissionsCreateType:
         if read:

From a0efe43887b305c5f96f6d2f9c24d7701c803a7f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 08:18:25 +0100
Subject: [PATCH 44/99] refactor: format text analyzer configuration for better
 readability

---
 integration/test_collection_config.py  | 12 +++++++++---
 test/collection/test_config.py         |  8 ++++++--
 weaviate/collections/classes/config.py |  1 -
 3 files changed, 15 insertions(+), 6 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index a125a3ef2..38ed3f141 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2572,7 +2572,9 @@ def test_property_text_analyzer_ascii_fold_immutable(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["é"]),
+                text_analyzer=Configure.TextAnalyzer.custom(
+                    ascii_fold=True, ascii_fold_ignore=["é"]
+                ),
             ),
         ],
     )
@@ -2634,7 +2636,9 @@ def test_property_ascii_fold_across_tokenizations(
                 name="prop",
                 data_type=DataType.TEXT,
                 tokenization=tokenization,
-                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["é"]),
+                text_analyzer=Configure.TextAnalyzer.custom(
+                    ascii_fold=True, ascii_fold_ignore=["é"]
+                ),
             ),
             Property(
                 name="prop_no_ignore",
@@ -2752,7 +2756,9 @@ def test_property_ascii_fold_with_filters(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["é"]),
+                text_analyzer=Configure.TextAnalyzer.custom(
+                    ascii_fold=True, ascii_fold_ignore=["é"]
+                ),
             ),
         ],
     )
diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index 67e4f6da3..5c468d5ea 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -3044,7 +3044,9 @@ def test_property_with_ascii_fold_and_ignore(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["é", "ñ"]),
+            text_analyzer=Configure.TextAnalyzer.custom(
+                ascii_fold=True, ascii_fold_ignore=["é", "ñ"]
+            ),
         )
         out = prop._to_dict()
         assert out["textAnalyzer"] == {
@@ -3065,7 +3067,9 @@ def test_nested_property_with_text_analyzer(self) -> None:
                 Property(
                     name="title",
                     data_type=DataType.TEXT,
-                    text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["ñ"]),
+                    text_analyzer=Configure.TextAnalyzer.custom(
+                        ascii_fold=True, ascii_fold_ignore=["ñ"]
+                    ),
                 ),
             ],
         )
diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 3e6867985..a4af8aadd 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -2263,7 +2263,6 @@ def custom(
         )
 
 
-
 class Property(_ConfigCreateModel):
     """This class defines the structure of a data property that a collection can have within Weaviate.
 

From fa92fc28f23078baced7fb01f8cbb49e9ecf69da Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 08:22:26 +0100
Subject: [PATCH 45/99] refactor: remove server side behavior tests

---
 integration/test_collection_config.py | 277 +-------------------------
 1 file changed, 9 insertions(+), 268 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 38ed3f141..285c09fe2 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -41,7 +41,6 @@
     _VectorizerConfigCreate,
     IndexName,
 )
-from weaviate.collections.classes.filters import Filter
 from weaviate.collections.classes.tenants import Tenant
 from weaviate.exceptions import (
     UnexpectedStatusCodeError,
@@ -2276,34 +2275,9 @@ def test_collection_stopword_presets(collection_factory: CollectionFactory) -> N
     assert title_en.text_analyzer.stopword_preset == "en"
     assert plain.text_analyzer is None
 
-    collection.data.insert(
-        {
-            "title_fr": "le chat noir",
-            "title_en": "the black cat",
-            "plain": "the black cat",
-        }
-    )
-
-    # title_fr filters 'le' (in user preset) → BM25 for 'le' yields no match,
-    # but 'chat' still matches.
-    res = collection.query.bm25(query="le", query_properties=["title_fr"])
-    assert len(res.objects) == 0
-    res = collection.query.bm25(query="chat", query_properties=["title_fr"])
-    assert len(res.objects) == 1
-
-    # title_en filters 'the' (built-in en) but matches 'cat'.
-    res = collection.query.bm25(query="the", query_properties=["title_en"])
-    assert len(res.objects) == 0
-    res = collection.query.bm25(query="cat", query_properties=["title_en"])
-    assert len(res.objects) == 1
-
 
 def test_collection_stopword_presets_update(collection_factory: CollectionFactory) -> None:
-    """Updating the contents of an existing stopword preset takes effect immediately.
-
-    Queries against properties that reference the preset see the new contents on the
-    next request.
-    """
+    """Updating a stopword preset is reflected in the config."""
     dummy = collection_factory("dummy")
     if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
         pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
@@ -2323,15 +2297,9 @@ def test_collection_stopword_presets_update(collection_factory: CollectionFactor
         ],
     )
 
-    collection.data.insert({"title_fr": "le chat et la souris"})
-
-    # Baseline: 'le' is filtered, 'la' is not (it isn't in the preset yet).
-    res = collection.query.bm25(query="le", query_properties=["title_fr"])
-    assert len(res.objects) == 0
-    res = collection.query.bm25(query="la", query_properties=["title_fr"])
-    assert len(res.objects) == 1
+    config = collection.config.get()
+    assert config.inverted_index_config.stopword_presets == {"fr": ["le"]}
 
-    # Replace the preset contents: drop 'le', add 'la'.
     collection.config.update(
         inverted_index_config=Reconfigure.inverted_index(
             stopword_presets={"fr": ["la"]},
@@ -2341,12 +2309,6 @@ def test_collection_stopword_presets_update(collection_factory: CollectionFactor
     config = collection.config.get()
     assert config.inverted_index_config.stopword_presets == {"fr": ["la"]}
 
-    # After update: 'le' now passes through, 'la' is filtered.
-    res = collection.query.bm25(query="le", query_properties=["title_fr"])
-    assert len(res.objects) == 1
-    res = collection.query.bm25(query="la", query_properties=["title_fr"])
-    assert len(res.objects) == 0
-
 
 def test_collection_stopword_presets_remove_in_use_is_rejected(
     collection_factory: CollectionFactory,
@@ -2476,7 +2438,7 @@ def test_collection_stopword_presets_remove_referenced_by_nested_property_is_rej
 def test_collection_user_defined_stopword_preset_overrides_builtin(
     collection_factory: CollectionFactory,
 ) -> None:
-    """A user-defined preset named 'en' replaces the built-in 'en' for properties of this collection."""
+    """A user-defined preset named 'en' is accepted and reflected in the config."""
     dummy = collection_factory("dummy")
     if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
         pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
@@ -2484,8 +2446,6 @@ def test_collection_user_defined_stopword_preset_overrides_builtin(
     collection = collection_factory(
         vectorizer_config=Configure.Vectorizer.none(),
         inverted_index_config=Configure.inverted_index(
-            # Shadow the built-in 'en' with a user-defined preset that filters
-            # 'hello' but NOT 'the'.
             stopword_presets={"en": ["hello"]},
         ),
         properties=[
@@ -2498,17 +2458,11 @@ def test_collection_user_defined_stopword_preset_overrides_builtin(
         ],
     )
 
-    collection.data.insert({"title": "the quick hello world"})
-
-    # 'the' is no longer filtered (built-in en is overridden) → matches.
-    res = collection.query.bm25(query="the", query_properties=["title"])
-    assert len(res.objects) == 1
-    # 'hello' was added by the override → filtered.
-    res = collection.query.bm25(query="hello", query_properties=["title"])
-    assert len(res.objects) == 0
-    # 'quick' is in neither list → matches.
-    res = collection.query.bm25(query="quick", query_properties=["title"])
-    assert len(res.objects) == 1
+    config = collection.config.get()
+    assert config.inverted_index_config.stopword_presets == {"en": ["hello"]}
+    title = next(p for p in config.properties if p.name == "title")
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.stopword_preset == "en"
 
 
 def test_property_text_analyzer_combined_ascii_fold_and_stopword_preset(
@@ -2540,18 +2494,6 @@ def test_property_text_analyzer_combined_ascii_fold_and_stopword_preset(
     assert title.text_analyzer.ascii_fold is True
     assert title.text_analyzer.stopword_preset == "en"
 
-    collection.data.insert({"title": "The école française"})
-
-    # 'the' is filtered by built-in en → no match.
-    res = collection.query.bm25(query="the", query_properties=["title"])
-    assert len(res.objects) == 0
-    # 'ecole' matches because ascii_fold folds 'école'.
-    res = collection.query.bm25(query="ecole", query_properties=["title"])
-    assert len(res.objects) == 1
-    # 'francaise' matches because ascii_fold folds 'française'.
-    res = collection.query.bm25(query="francaise", query_properties=["title"])
-    assert len(res.objects) == 1
-
 
 def test_property_text_analyzer_ascii_fold_immutable(
     collection_factory: CollectionFactory,
@@ -2607,207 +2549,6 @@ def test_property_text_analyzer_ascii_fold_immutable(
     assert title2.text_analyzer.ascii_fold_ignore == ["ñ"]
 
 
-@pytest.mark.parametrize(
-    "tokenization,query_match,query_no_match",
-    [
-        (Tokenization.WORD, "école", "ecole"),
-        (Tokenization.LOWERCASE, "l'école", "l'ecole"),
-        (Tokenization.WHITESPACE, "L'école", "L'ecole"),
-        (Tokenization.FIELD, "L'école est fermée", "L'ecole est fermee"),
-        (Tokenization.TRIGRAM, "éco", "eco"),
-    ],
-    ids=["word", "lowercase", "whitespace", "field", "trigram"],
-)
-def test_property_ascii_fold_across_tokenizations(
-    collection_factory: CollectionFactory,
-    tokenization: Tokenization,
-    query_match: str,
-    query_no_match: str,
-) -> None:
-    """ascii_fold + ignore list ['é'] preserves é under every supported tokenization."""
-    dummy = collection_factory("dummy")
-    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
-        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
-
-    collection = collection_factory(
-        vectorizer_config=Configure.Vectorizer.none(),
-        properties=[
-            Property(
-                name="prop",
-                data_type=DataType.TEXT,
-                tokenization=tokenization,
-                text_analyzer=Configure.TextAnalyzer.custom(
-                    ascii_fold=True, ascii_fold_ignore=["é"]
-                ),
-            ),
-            Property(
-                name="prop_no_ignore",
-                data_type=DataType.TEXT,
-                tokenization=tokenization,
-                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True),
-            ),
-        ],
-    )
-    collection.data.insert(
-        {
-            "prop": "L'école est fermée",
-            "prop_no_ignore": "L'école est fermée",
-        }
-    )
-
-    # On the property with the ignore list, the accented form matches but the
-    # ASCII-folded form does not.
-    res = collection.query.bm25(query=query_match, query_properties=["prop"])
-    assert len(res.objects) == 1
-    res = collection.query.bm25(query=query_no_match, query_properties=["prop"])
-    assert len(res.objects) == 0
-
-    # On the property without the ignore list, both forms match.
-    res = collection.query.bm25(query=query_match, query_properties=["prop_no_ignore"])
-    assert len(res.objects) == 1
-    res = collection.query.bm25(query=query_no_match, query_properties=["prop_no_ignore"])
-    assert len(res.objects) == 1
-
-
-def test_property_ascii_fold_multi_char_ignore(
-    collection_factory: CollectionFactory,
-) -> None:
-    """An ignore list with multiple characters preserves all of them.
-
-    Characters not in the ignore list are still folded normally.
-    """
-    dummy = collection_factory("dummy")
-    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
-        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
-
-    collection = collection_factory(
-        vectorizer_config=Configure.Vectorizer.none(),
-        properties=[
-            Property(
-                name="multi_ignore",
-                data_type=DataType.TEXT,
-                tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(
-                    ascii_fold=True,
-                    ascii_fold_ignore=["é", "ü", "ñ", "ø"],
-                ),
-            ),
-        ],
-    )
-    collection.data.insert_many(
-        [
-            {"multi_ignore": "résumé"},
-            {"multi_ignore": "über"},
-            {"multi_ignore": "El Niño"},
-            {"multi_ignore": "Ørsted"},
-            {"multi_ignore": "São Paulo"},
-            {"multi_ignore": "naïve café"},
-        ]
-    )
-
-    # All four ignored chars are preserved → ASCII-folded form does not match,
-    # but the accented form does.
-    for accented, folded in [
-        ("résumé", "resume"),
-        ("über", "uber"),
-        ("niño", "nino"),
-        ("ørsted", "orsted"),
-    ]:
-        match = collection.query.bm25(query=accented, query_properties=["multi_ignore"])
-        no_match = collection.query.bm25(query=folded, query_properties=["multi_ignore"])
-        assert len(match.objects) == 1, f"{accented} should match"
-        assert len(no_match.objects) == 0, f"{folded} should not match (char in ignore list)"
-
-    # Non-ignored accents (ã, ï) still fold normally.
-    res = collection.query.bm25(query="sao", query_properties=["multi_ignore"])
-    assert len(res.objects) == 1
-    res = collection.query.bm25(query="naive", query_properties=["multi_ignore"])
-    assert len(res.objects) == 1
-
-    # The "naïve café" doc has both an ignored char (é) and a non-ignored one (ï):
-    # 'naive' matches (ï folded), 'cafe' does not (é preserved).
-    res = collection.query.fetch_objects(
-        filters=Filter.by_property("multi_ignore").equal("naive café"),
-        limit=10,
-    )
-    assert len(res.objects) == 1
-
-
-def test_property_ascii_fold_with_filters(
-    collection_factory: CollectionFactory,
-) -> None:
-    """Equal and Like filters respect ascii_fold and ascii_fold_ignore."""
-    dummy = collection_factory("dummy")
-    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
-        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
-
-    collection = collection_factory(
-        vectorizer_config=Configure.Vectorizer.none(),
-        properties=[
-            # Folds everything.
-            Property(
-                name="body",
-                data_type=DataType.TEXT,
-                tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True),
-            ),
-            # Folds everything except 'é'.
-            Property(
-                name="title",
-                data_type=DataType.TEXT,
-                tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(
-                    ascii_fold=True, ascii_fold_ignore=["é"]
-                ),
-            ),
-        ],
-    )
-    collection.data.insert_many(
-        [
-            {"title": "L'école est fermée", "body": "L'école est fermée"},
-            {"title": "cafe résumé", "body": "cafe résumé"},
-        ]
-    )
-
-    # Equal on body (full fold): 'ecole' matches.
-    res = collection.query.fetch_objects(
-        filters=Filter.by_property("body").equal("ecole"),
-        limit=5,
-    )
-    assert len(res.objects) == 1
-
-    # Equal on title (é preserved): 'ecole' does not, but 'école' does.
-    res = collection.query.fetch_objects(
-        filters=Filter.by_property("title").equal("école"),
-        limit=5,
-    )
-    assert len(res.objects) == 1
-    res = collection.query.fetch_objects(
-        filters=Filter.by_property("title").equal("ecole"),
-        limit=5,
-    )
-    assert len(res.objects) == 0
-
-    # Like on body (full fold): 'ecol*' matches.
-    res = collection.query.fetch_objects(
-        filters=Filter.by_property("body").like("ecol*"),
-        limit=5,
-    )
-    assert len(res.objects) == 1
-
-    # Like on title (é preserved): 'écol*' matches, 'ecol*' does not.
-    res = collection.query.fetch_objects(
-        filters=Filter.by_property("title").like("écol*"),
-        limit=5,
-    )
-    assert len(res.objects) == 1
-    res = collection.query.fetch_objects(
-        filters=Filter.by_property("title").like("ecol*"),
-        limit=5,
-    )
-    assert len(res.objects) == 0
-
-
 def test_stopwords_roundtrip_from_dict(collection_factory: CollectionFactory) -> None:
     collection = collection_factory(
         inverted_index_config=Configure.inverted_index(

From 27cd0a4ddd1a5dd34c1629083b7461930abf7c54 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 08:25:19 +0100
Subject: [PATCH 46/99] test: add stopword presets roundtrip tests for Weaviate
 collections

---
 integration/test_collection_config.py | 80 +++++++++++++++++++++++++++
 1 file changed, 80 insertions(+)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 285c09fe2..39c40a366 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2570,3 +2570,83 @@ def test_stopwords_roundtrip_from_dict(collection_factory: CollectionFactory) ->
         assert config == new
         assert config.to_dict() == new.to_dict()
         client.collections.delete(name)
+
+
+def test_stopword_presets_roundtrip_from_dict(
+    collection_factory: CollectionFactory,
+) -> None:
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("stopword_presets requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        inverted_index_config=Configure.inverted_index(
+            stopword_presets={"fr": ["le", "la", "les"]},
+        ),
+        properties=[
+            Property(
+                name="title",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
+            ),
+        ],
+    )
+
+    config = collection.config.get()
+    assert config.inverted_index_config.stopword_presets == {"fr": ["le", "la", "les"]}
+    title = next(p for p in config.properties if p.name == "title")
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.stopword_preset == "fr"
+
+    name = f"TestPresetRoundtrip{collection.name}"
+    config.name = name
+    with weaviate.connect_to_local() as client:
+        client.collections.delete(name)
+        client.collections.create_from_dict(config.to_dict())
+        new = client.collections.use(name).config.get()
+        assert config == new
+        assert config.to_dict() == new.to_dict()
+        client.collections.delete(name)
+
+
+def test_text_analyzer_roundtrip_from_dict(
+    collection_factory: CollectionFactory,
+) -> None:
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
+
+    collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
+        properties=[
+            Property(
+                name="title",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=Configure.TextAnalyzer.custom(
+                    ascii_fold=True,
+                    ascii_fold_ignore=["é"],
+                    stopword_preset=StopwordsPreset.EN,
+                ),
+            ),
+        ],
+    )
+
+    config = collection.config.get()
+    title = next(p for p in config.properties if p.name == "title")
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.ascii_fold is True
+    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+    assert title.text_analyzer.stopword_preset == "en"
+
+    name = f"TestAnalyzerRoundtrip{collection.name}"
+    config.name = name
+    with weaviate.connect_to_local() as client:
+        client.collections.delete(name)
+        client.collections.create_from_dict(config.to_dict())
+        new = client.collections.use(name).config.get()
+        assert config == new
+        assert config.to_dict() == new.to_dict()
+        client.collections.delete(name)

From a241d8c6343b246b93f76b0913ade5b934b78df4 Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Tue, 14 Apr 2026 09:30:52 +0200
Subject: [PATCH 47/99] Fix formatting

---
 integration/test_rbac.py | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/integration/test_rbac.py b/integration/test_rbac.py
index 53206abac..86719b6ce 100644
--- a/integration/test_rbac.py
+++ b/integration/test_rbac.py
@@ -437,9 +437,7 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
-                mcp_permissions=[
-                    MCPPermissionOutput(actions={Actions.MCP.MANAGE})
-                ],
+                mcp_permissions=[MCPPermissionOutput(actions={Actions.MCP.MANAGE})],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],

From 83c2431e151bed0d8225aaf9bf7b553650c3a502 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 08:31:59 +0100
Subject: [PATCH 48/99] refactor: remove unnecessary stopword preset coercion
 from _TextAnalyzerConfigCreate

---
 weaviate/collections/classes/config.py | 11 ++---------
 1 file changed, 2 insertions(+), 9 deletions(-)

diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index a4af8aadd..476fb302e 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -2216,15 +2216,6 @@ class _TextAnalyzerConfigCreate(_ConfigCreateModel):
         default=None, alias="stopword_preset"
     )
 
-    @field_validator("stopwordPreset", mode="before")
-    @classmethod
-    def _coerce_stopword_preset(cls, v: Any) -> Any:
-        # Pydantic preserves the StopwordsPreset enum instance through model_dump,
-        # but the wire format must be a plain string. Coerce at construction time.
-        if isinstance(v, StopwordsPreset):
-            return v.value
-        return v
-
     @model_validator(mode="after")
     def _validate_ascii_fold_ignore(self) -> "_TextAnalyzerConfigCreate":
         if self.asciiFold is not True and self.asciiFoldIgnore is not None:
@@ -2324,6 +2315,8 @@ def _to_dict(
                 if isinstance(self.nestedProperties, list)
                 else [self.nestedProperties._to_dict()]
             )
+        if self.textAnalyzer is not None:
+            ret_dict["textAnalyzer"] = self.textAnalyzer._to_dict()
         return ret_dict
 
 

From 4e0a0f20503dcfead5417e1fbc6686070b860c41 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 09:11:04 +0100
Subject: [PATCH 49/99] refactor: replace custom text analyzer method with a
 direct function call

---
 integration/test_collection_config.py  | 28 +++++++-------
 test/collection/test_config.py         | 18 ++++-----
 weaviate/collections/classes/config.py | 51 +++++++++++---------------
 3 files changed, 42 insertions(+), 55 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 39c40a366..527bad130 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2218,7 +2218,7 @@ def test_property_text_analyzer_ascii_fold_version_gate(
                     name="title",
                     data_type=DataType.TEXT,
                     tokenization=Tokenization.WORD,
-                    text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True),
+                    text_analyzer=Configure.TextAnalyzer(ascii_fold=True),
                 ),
             ],
         )
@@ -2245,14 +2245,14 @@ def test_collection_stopword_presets(collection_factory: CollectionFactory) -> N
                 name="title_fr",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
+                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
             ),
             # Built-in English preset, set per property.
             Property(
                 name="title_en",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset=StopwordsPreset.EN),
+                text_analyzer=Configure.TextAnalyzer(stopword_preset=StopwordsPreset.EN),
             ),
             # No stopword override → uses the collection-level default.
             Property(
@@ -2292,7 +2292,7 @@ def test_collection_stopword_presets_update(collection_factory: CollectionFactor
                 name="title_fr",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
+                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
             ),
         ],
     )
@@ -2328,7 +2328,7 @@ def test_collection_stopword_presets_remove_in_use_is_rejected(
                 name="title_fr",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
+                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
             ),
         ],
     )
@@ -2381,7 +2381,7 @@ def test_collection_stopword_presets_remove_unused_is_allowed(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
+                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
             ),
         ],
     )
@@ -2418,7 +2418,7 @@ def test_collection_stopword_presets_remove_referenced_by_nested_property_is_rej
                         name="body",
                         data_type=DataType.TEXT,
                         tokenization=Tokenization.WORD,
-                        text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
+                        text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
                     ),
                 ],
             ),
@@ -2453,7 +2453,7 @@ def test_collection_user_defined_stopword_preset_overrides_builtin(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="en"),
+                text_analyzer=Configure.TextAnalyzer(stopword_preset="en"),
             ),
         ],
     )
@@ -2480,7 +2480,7 @@ def test_property_text_analyzer_combined_ascii_fold_and_stopword_preset(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(
+                text_analyzer=Configure.TextAnalyzer(
                     ascii_fold=True,
                     stopword_preset=StopwordsPreset.EN,
                 ),
@@ -2514,9 +2514,7 @@ def test_property_text_analyzer_ascii_fold_immutable(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(
-                    ascii_fold=True, ascii_fold_ignore=["é"]
-                ),
+                text_analyzer=Configure.TextAnalyzer(ascii_fold=True, ascii_fold_ignore=["é"]),
             ),
         ],
     )
@@ -2535,7 +2533,7 @@ def test_property_text_analyzer_ascii_fold_immutable(
             name="title2",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True, ascii_fold_ignore=["ñ"]),
+            text_analyzer=Configure.TextAnalyzer(ascii_fold=True, ascii_fold_ignore=["ñ"]),
         ),
     )
     config = collection.config.get()
@@ -2589,7 +2587,7 @@ def test_stopword_presets_roundtrip_from_dict(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
+                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
             ),
         ],
     )
@@ -2625,7 +2623,7 @@ def test_text_analyzer_roundtrip_from_dict(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer.custom(
+                text_analyzer=Configure.TextAnalyzer(
                     ascii_fold=True,
                     ascii_fold_ignore=["é"],
                     stopword_preset=StopwordsPreset.EN,
diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index 5c468d5ea..f06144ddb 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -3035,7 +3035,7 @@ def test_property_with_ascii_fold_only(self) -> None:
         prop = Property(
             name="title",
             data_type=DataType.TEXT,
-            text_analyzer=Configure.TextAnalyzer.custom(ascii_fold=True),
+            text_analyzer=Configure.TextAnalyzer(ascii_fold=True),
         )
         assert prop._to_dict()["textAnalyzer"] == {"asciiFold": True}
 
@@ -3044,9 +3044,7 @@ def test_property_with_ascii_fold_and_ignore(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer.custom(
-                ascii_fold=True, ascii_fold_ignore=["é", "ñ"]
-            ),
+            text_analyzer=Configure.TextAnalyzer(ascii_fold=True, ascii_fold_ignore=["é", "ñ"]),
         )
         out = prop._to_dict()
         assert out["textAnalyzer"] == {
@@ -3067,9 +3065,7 @@ def test_nested_property_with_text_analyzer(self) -> None:
                 Property(
                     name="title",
                     data_type=DataType.TEXT,
-                    text_analyzer=Configure.TextAnalyzer.custom(
-                        ascii_fold=True, ascii_fold_ignore=["ñ"]
-                    ),
+                    text_analyzer=Configure.TextAnalyzer(ascii_fold=True, ascii_fold_ignore=["ñ"]),
                 ),
             ],
         )
@@ -3090,7 +3086,7 @@ def test_text_analyzer_stopword_preset_builtin_enum(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer.custom(stopword_preset=StopwordsPreset.EN),
+            text_analyzer=Configure.TextAnalyzer(stopword_preset=StopwordsPreset.EN),
         )
         assert prop._to_dict()["textAnalyzer"] == {"stopwordPreset": "en"}
 
@@ -3099,7 +3095,7 @@ def test_text_analyzer_stopword_preset_user_defined_string(self) -> None:
             name="title_fr",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
+            text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
         )
         assert prop._to_dict()["textAnalyzer"] == {"stopwordPreset": "fr"}
 
@@ -3108,7 +3104,7 @@ def test_text_analyzer_combined_ascii_fold_and_stopword_preset(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer.custom(
+            text_analyzer=Configure.TextAnalyzer(
                 ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset="fr"
             ),
         )
@@ -3123,7 +3119,7 @@ def test_text_analyzer_stopword_preset_only_omits_other_keys(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer.custom(stopword_preset="fr"),
+            text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
         )
         out = prop._to_dict()
         assert "asciiFold" not in out["textAnalyzer"]
diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 476fb302e..d8c06cdd1 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -2223,35 +2223,28 @@ def _validate_ascii_fold_ignore(self) -> "_TextAnalyzerConfigCreate":
         return self
 
 
-class _TextAnalyzer:
-    """Factory class for creating text analyzer configurations.
-
-    Use ``Configure.TextAnalyzer`` to access these methods.
+def _text_analyzer(
+    ascii_fold: Optional[bool] = None,
+    ascii_fold_ignore: Optional[List[str]] = None,
+    stopword_preset: Optional[Union[StopwordsPreset, str]] = None,
+) -> _TextAnalyzerConfigCreate:
+    """Create a text analyzer config for a property.
+
+    Args:
+        ascii_fold: If True, accent/diacritic marks are folded to their base
+            characters during indexing and search (e.g. 'école' matches 'ecole').
+        ascii_fold_ignore: Optional list of characters that should be excluded
+            from ASCII folding (e.g. ``['é']`` keeps 'é' from being folded to
+            'e'). Requires ``ascii_fold=True``.
+        stopword_preset: Stopword preset name to override the collection-level
+            stopwords for this property. Accepts a ``StopwordsPreset`` or a
+            user-defined preset name.
     """
-
-    @staticmethod
-    def custom(
-        ascii_fold: Optional[bool] = None,
-        ascii_fold_ignore: Optional[List[str]] = None,
-        stopword_preset: Optional[Union[StopwordsPreset, str]] = None,
-    ) -> _TextAnalyzerConfigCreate:
-        """Create a text analyzer config with custom settings.
-
-        Args:
-            ascii_fold: If True, accent/diacritic marks are folded to their base
-                characters during indexing and search (e.g. 'école' matches 'ecole').
-            ascii_fold_ignore: Optional list of characters that should be excluded
-                from ASCII folding (e.g. ``['é']`` keeps 'é' from being folded to
-                'e'). Requires ``ascii_fold=True``.
-            stopword_preset: Stopword preset name to override the collection-level
-                stopwords for this property. Accepts a ``StopwordsPreset`` or a
-                user-defined preset name.
-        """
-        return _TextAnalyzerConfigCreate(
-            ascii_fold=ascii_fold,
-            ascii_fold_ignore=ascii_fold_ignore,
-            stopword_preset=stopword_preset,
-        )
+    return _TextAnalyzerConfigCreate(
+        ascii_fold=ascii_fold,
+        ascii_fold_ignore=ascii_fold_ignore,
+        stopword_preset=stopword_preset,
+    )
 
 
 class Property(_ConfigCreateModel):
@@ -2665,7 +2658,7 @@ class Configure:
     MultiVectors = _MultiVectors
     ObjectTTL = _ObjectTTL
     Replication = _Replication
-    TextAnalyzer = _TextAnalyzer
+    TextAnalyzer = staticmethod(_text_analyzer)
 
     @staticmethod
     def inverted_index(

From 38c7f44c9febfb7aaabe6c04ad4181079a44ce77 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 09:45:19 +0100
Subject: [PATCH 50/99] chore: remove unused deprecated import from config.py

---
 weaviate/collections/classes/config.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 624baf9dd..5ea64e183 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -14,7 +14,6 @@
     cast,
 )
 
-from deprecation import deprecated as docstring_deprecated
 from pydantic import (
     AnyHttpUrl,
     Field,

From b3eb0ac219d38e0bc3419194840e1016ab169a6c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 12:24:10 +0100
Subject: [PATCH 51/99] chore: update WEAVIATE_137 version to
 1.37.0-rc.1-578c4eb in workflow

---
 .github/workflows/main.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index 8ad6083a0..d3dfd1fa4 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -28,7 +28,7 @@ env:
   WEAVIATE_134: 1.34.19
   WEAVIATE_135: 1.35.16-efdedfa
   WEAVIATE_136: 1.36.9-d905e6c
-  WEAVIATE_137: 1.37.0-rc.0-b313954.amd64
+  WEAVIATE_137: 1.37.0-rc.1-578c4eb
 
 jobs:
   lint-and-format:

From ceef2712fe2e1b2e637129b4ef45fbe0ecafd6c9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 12:27:22 +0100
Subject: [PATCH 52/99] refactor: update text analyzer method to use new static
 method in Configure

---
 integration/test_collection_config.py  | 26 +++++++-------
 test/collection/test_config.py         | 14 ++++----
 weaviate/collections/classes/config.py | 49 +++++++++++++-------------
 3 files changed, 44 insertions(+), 45 deletions(-)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 25f04e9ae..45fa6ab17 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2222,7 +2222,7 @@ def test_property_text_analyzer_ascii_fold_version_gate(
                     name="title",
                     data_type=DataType.TEXT,
                     tokenization=Tokenization.WORD,
-                    text_analyzer=Configure.TextAnalyzer(ascii_fold=True),
+                    text_analyzer=Configure.text_analyzer(ascii_fold=True),
                 ),
             ],
         )
@@ -2249,14 +2249,14 @@ def test_collection_stopword_presets(collection_factory: CollectionFactory) -> N
                 name="title_fr",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
+                text_analyzer=Configure.text_analyzer(stopword_preset="fr"),
             ),
             # Built-in English preset, set per property.
             Property(
                 name="title_en",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(stopword_preset=StopwordsPreset.EN),
+                text_analyzer=Configure.text_analyzer(stopword_preset=StopwordsPreset.EN),
             ),
             # No stopword override → uses the collection-level default.
             Property(
@@ -2296,7 +2296,7 @@ def test_collection_stopword_presets_update(collection_factory: CollectionFactor
                 name="title_fr",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
+                text_analyzer=Configure.text_analyzer(stopword_preset="fr"),
             ),
         ],
     )
@@ -2332,7 +2332,7 @@ def test_collection_stopword_presets_remove_in_use_is_rejected(
                 name="title_fr",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
+                text_analyzer=Configure.text_analyzer(stopword_preset="fr"),
             ),
         ],
     )
@@ -2385,7 +2385,7 @@ def test_collection_stopword_presets_remove_unused_is_allowed(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
+                text_analyzer=Configure.text_analyzer(stopword_preset="fr"),
             ),
         ],
     )
@@ -2422,7 +2422,7 @@ def test_collection_stopword_presets_remove_referenced_by_nested_property_is_rej
                         name="body",
                         data_type=DataType.TEXT,
                         tokenization=Tokenization.WORD,
-                        text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
+                        text_analyzer=Configure.text_analyzer(stopword_preset="fr"),
                     ),
                 ],
             ),
@@ -2457,7 +2457,7 @@ def test_collection_user_defined_stopword_preset_overrides_builtin(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(stopword_preset="en"),
+                text_analyzer=Configure.text_analyzer(stopword_preset="en"),
             ),
         ],
     )
@@ -2484,7 +2484,7 @@ def test_property_text_analyzer_combined_ascii_fold_and_stopword_preset(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(
+                text_analyzer=Configure.text_analyzer(
                     ascii_fold=True,
                     stopword_preset=StopwordsPreset.EN,
                 ),
@@ -2518,7 +2518,7 @@ def test_property_text_analyzer_ascii_fold_immutable(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(ascii_fold=True, ascii_fold_ignore=["é"]),
+                text_analyzer=Configure.text_analyzer(ascii_fold=True, ascii_fold_ignore=["é"]),
             ),
         ],
     )
@@ -2537,7 +2537,7 @@ def test_property_text_analyzer_ascii_fold_immutable(
             name="title2",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer(ascii_fold=True, ascii_fold_ignore=["ñ"]),
+            text_analyzer=Configure.text_analyzer(ascii_fold=True, ascii_fold_ignore=["ñ"]),
         ),
     )
     config = collection.config.get()
@@ -2591,7 +2591,7 @@ def test_stopword_presets_roundtrip_from_dict(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
+                text_analyzer=Configure.text_analyzer(stopword_preset="fr"),
             ),
         ],
     )
@@ -2627,7 +2627,7 @@ def test_text_analyzer_roundtrip_from_dict(
                 name="title",
                 data_type=DataType.TEXT,
                 tokenization=Tokenization.WORD,
-                text_analyzer=Configure.TextAnalyzer(
+                text_analyzer=Configure.text_analyzer(
                     ascii_fold=True,
                     ascii_fold_ignore=["é"],
                     stopword_preset=StopwordsPreset.EN,
diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index f06144ddb..5a69b21c5 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -3035,7 +3035,7 @@ def test_property_with_ascii_fold_only(self) -> None:
         prop = Property(
             name="title",
             data_type=DataType.TEXT,
-            text_analyzer=Configure.TextAnalyzer(ascii_fold=True),
+            text_analyzer=Configure.text_analyzer(ascii_fold=True),
         )
         assert prop._to_dict()["textAnalyzer"] == {"asciiFold": True}
 
@@ -3044,7 +3044,7 @@ def test_property_with_ascii_fold_and_ignore(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer(ascii_fold=True, ascii_fold_ignore=["é", "ñ"]),
+            text_analyzer=Configure.text_analyzer(ascii_fold=True, ascii_fold_ignore=["é", "ñ"]),
         )
         out = prop._to_dict()
         assert out["textAnalyzer"] == {
@@ -3065,7 +3065,7 @@ def test_nested_property_with_text_analyzer(self) -> None:
                 Property(
                     name="title",
                     data_type=DataType.TEXT,
-                    text_analyzer=Configure.TextAnalyzer(ascii_fold=True, ascii_fold_ignore=["ñ"]),
+                    text_analyzer=Configure.text_analyzer(ascii_fold=True, ascii_fold_ignore=["ñ"]),
                 ),
             ],
         )
@@ -3086,7 +3086,7 @@ def test_text_analyzer_stopword_preset_builtin_enum(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer(stopword_preset=StopwordsPreset.EN),
+            text_analyzer=Configure.text_analyzer(stopword_preset=StopwordsPreset.EN),
         )
         assert prop._to_dict()["textAnalyzer"] == {"stopwordPreset": "en"}
 
@@ -3095,7 +3095,7 @@ def test_text_analyzer_stopword_preset_user_defined_string(self) -> None:
             name="title_fr",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
+            text_analyzer=Configure.text_analyzer(stopword_preset="fr"),
         )
         assert prop._to_dict()["textAnalyzer"] == {"stopwordPreset": "fr"}
 
@@ -3104,7 +3104,7 @@ def test_text_analyzer_combined_ascii_fold_and_stopword_preset(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer(
+            text_analyzer=Configure.text_analyzer(
                 ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset="fr"
             ),
         )
@@ -3119,7 +3119,7 @@ def test_text_analyzer_stopword_preset_only_omits_other_keys(self) -> None:
             name="title",
             data_type=DataType.TEXT,
             tokenization=Tokenization.WORD,
-            text_analyzer=Configure.TextAnalyzer(stopword_preset="fr"),
+            text_analyzer=Configure.text_analyzer(stopword_preset="fr"),
         )
         out = prop._to_dict()
         assert "asciiFold" not in out["textAnalyzer"]
diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index f746dfd03..7874cd8e2 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -2222,30 +2222,6 @@ def _validate_ascii_fold_ignore(self) -> "_TextAnalyzerConfigCreate":
         return self
 
 
-def _text_analyzer(
-    ascii_fold: Optional[bool] = None,
-    ascii_fold_ignore: Optional[List[str]] = None,
-    stopword_preset: Optional[Union[StopwordsPreset, str]] = None,
-) -> _TextAnalyzerConfigCreate:
-    """Create a text analyzer config for a property.
-
-    Args:
-        ascii_fold: If True, accent/diacritic marks are folded to their base
-            characters during indexing and search (e.g. 'école' matches 'ecole').
-        ascii_fold_ignore: Optional list of characters that should be excluded
-            from ASCII folding (e.g. ``['é']`` keeps 'é' from being folded to
-            'e'). Requires ``ascii_fold=True``.
-        stopword_preset: Stopword preset name to override the collection-level
-            stopwords for this property. Accepts a ``StopwordsPreset`` or a
-            user-defined preset name.
-    """
-    return _TextAnalyzerConfigCreate(
-        ascii_fold=ascii_fold,
-        ascii_fold_ignore=ascii_fold_ignore,
-        stopword_preset=stopword_preset,
-    )
-
-
 class Property(_ConfigCreateModel):
     """This class defines the structure of a data property that a collection can have within Weaviate.
 
@@ -2657,7 +2633,30 @@ class Configure:
     MultiVectors = _MultiVectors
     ObjectTTL = _ObjectTTL
     Replication = _Replication
-    TextAnalyzer = staticmethod(_text_analyzer)
+
+    @staticmethod
+    def text_analyzer(
+        ascii_fold: Optional[bool] = None,
+        ascii_fold_ignore: Optional[List[str]] = None,
+        stopword_preset: Optional[Union[StopwordsPreset, str]] = None,
+    ) -> _TextAnalyzerConfigCreate:
+        """Create a text analyzer config for a property.
+
+        Args:
+            ascii_fold: If True, accent/diacritic marks are folded to their base
+                characters during indexing and search (e.g. 'école' matches 'ecole').
+            ascii_fold_ignore: Optional list of characters that should be excluded
+                from ASCII folding (e.g. ``['é']`` keeps 'é' from being folded to
+                'e'). Requires ``ascii_fold=True``.
+            stopword_preset: Stopword preset name to override the collection-level
+                stopwords for this property. Accepts a ``StopwordsPreset`` or a
+                user-defined preset name.
+        """
+        return _TextAnalyzerConfigCreate(
+            ascii_fold=ascii_fold,
+            ascii_fold_ignore=ascii_fold_ignore,
+            stopword_preset=stopword_preset,
+        )
 
     @staticmethod
     def inverted_index(

From 5e751bfbbffaa236c083ec1cfac00815f9962701 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 12:29:50 +0100
Subject: [PATCH 53/99] test: add stopwords roundtrip test with ASCII folding
 configuration

---
 integration/test_collection_config.py | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/integration/test_collection_config.py b/integration/test_collection_config.py
index 45fa6ab17..3f33a30b7 100644
--- a/integration/test_collection_config.py
+++ b/integration/test_collection_config.py
@@ -2552,16 +2552,38 @@ def test_property_text_analyzer_ascii_fold_immutable(
 
 
 def test_stopwords_roundtrip_from_dict(collection_factory: CollectionFactory) -> None:
+    dummy = collection_factory("dummy")
+    if dummy._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("text_analyzer requires Weaviate >= 1.37.0")
+
     collection = collection_factory(
+        vectorizer_config=Configure.Vectorizer.none(),
         inverted_index_config=Configure.inverted_index(
             stopwords_additions=["a"],
             stopwords_preset=StopwordsPreset.EN,
             stopwords_removals=["the"],
+            stopword_presets={"fr": ["le", "la", "les"]},
         ),
+        properties=[
+            Property(
+                name="title",
+                data_type=DataType.TEXT,
+                tokenization=Tokenization.WORD,
+                text_analyzer=Configure.text_analyzer(
+                    ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset="fr"
+                ),
+            ),
+        ],
     )
     config = collection.config.get()
     assert config.inverted_index_config.stopwords.preset == StopwordsPreset.EN
     assert config.inverted_index_config.stopwords.removals == ["the"]
+    assert config.inverted_index_config.stopword_presets == {"fr": ["le", "la", "les"]}
+    title = next(p for p in config.properties if p.name == "title")
+    assert title.text_analyzer is not None
+    assert title.text_analyzer.ascii_fold is True
+    assert title.text_analyzer.ascii_fold_ignore == ["é"]
+    assert title.text_analyzer.stopword_preset == "fr"
 
     name = f"TestStopwordsRoundtrip{collection.name}"
     config.name = name

From 9c4295b5402c17ffad6d954a125225c3ac8f4370 Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Tue, 14 Apr 2026 16:01:06 +0200
Subject: [PATCH 54/99] Add query profiling

---
 integration/test_collection_query_profile.py  | 217 ++++++++++++++++++
 weaviate/collections/classes/grpc.py          |   4 +
 weaviate/collections/classes/internal.py      |  29 +++
 weaviate/collections/grpc/query.py            |   1 +
 weaviate/collections/queries/base_executor.py |  42 +++-
 weaviate/outputs/query.py                     |   6 +
 weaviate/proto/v1/v4216/v1/search_get_pb2.py  |  62 +++--
 weaviate/proto/v1/v4216/v1/search_get_pb2.pyi |  46 +++-
 weaviate/proto/v1/v5261/v1/search_get_pb2.py  |  62 +++--
 weaviate/proto/v1/v5261/v1/search_get_pb2.pyi |  46 +++-
 weaviate/proto/v1/v6300/v1/search_get_pb2.py  |  62 +++--
 weaviate/proto/v1/v6300/v1/search_get_pb2.pyi |  46 +++-
 12 files changed, 536 insertions(+), 87 deletions(-)
 create mode 100644 integration/test_collection_query_profile.py

diff --git a/integration/test_collection_query_profile.py b/integration/test_collection_query_profile.py
new file mode 100644
index 000000000..021c83802
--- /dev/null
+++ b/integration/test_collection_query_profile.py
@@ -0,0 +1,217 @@
+import re
+
+import pytest
+
+import weaviate
+from weaviate.collections.classes.config import Configure, DataType, Property
+from weaviate.collections.classes.data import DataObject
+from weaviate.collections.classes.grpc import MetadataQuery
+from weaviate.collections.classes.internal import SearchProfileReturn
+
+GO_DURATION_RE = re.compile(r"[\d.]+(ns|µs|ms|s|m|h)")
+
+
+def assert_go_duration(value: str, label: str = "") -> None:
+    """Assert that a string looks like a Go duration (e.g. '1.234ms', '5.458µs')."""
+    assert GO_DURATION_RE.fullmatch(
+        value
+    ), f"Expected Go duration format for {label!r}, got {value!r}"
+
+
+def assert_common_profile(profile: SearchProfileReturn) -> None:
+    """Assertions shared by every search profile regardless of type."""
+    assert len(profile.details) > 0, "Profile details should not be empty"
+    assert "total_took" in profile.details
+    assert_go_duration(profile.details["total_took"], "total_took")
+    for key, value in profile.details.items():
+        assert isinstance(key, str) and key != ""
+        assert isinstance(value, str) and value != ""
+
+
+@pytest.fixture(scope="module")
+def client():
+    client = weaviate.connect_to_local()
+    yield client
+    client.close()
+
+
+@pytest.fixture(scope="module")
+def collection_with_data(client: weaviate.WeaviateClient):
+    name = "TestQueryProfile"
+    client.collections.delete(name)
+    collection = client.collections.create(
+        name=name,
+        vectorizer_config=Configure.Vectorizer.none(),
+        properties=[
+            Property(name="text", data_type=DataType.TEXT),
+        ],
+    )
+    collection.data.insert_many(
+        [
+            DataObject(properties={"text": "hello world"}, vector=[1.0, 0.0, 0.0]),
+            DataObject(properties={"text": "goodbye world"}, vector=[0.0, 1.0, 0.0]),
+            DataObject(properties={"text": "foo bar baz"}, vector=[0.0, 0.0, 1.0]),
+        ]
+    )
+    yield collection
+    client.collections.delete(name)
+
+
+def test_fetch_objects_with_query_profile(collection_with_data):
+    """Test that query profiling works with fetch_objects (object lookup)."""
+    result = collection_with_data.query.fetch_objects(
+        return_metadata=MetadataQuery(query_profile=True),
+    )
+    assert len(result.objects) == 3
+    assert result.query_profile is not None
+    assert len(result.query_profile.shards) > 0
+
+    shard = result.query_profile.shards[0]
+    assert shard.name != ""
+    assert shard.node != ""
+
+    assert "object" in shard.searches
+    assert_common_profile(shard.searches["object"])
+
+
+def test_near_vector_with_query_profile(collection_with_data):
+    """Test that query profiling works with near_vector search."""
+    result = collection_with_data.query.near_vector(
+        near_vector=[1.0, 0.0, 0.0],
+        return_metadata=MetadataQuery(query_profile=True, distance=True),
+        limit=2,
+    )
+    assert len(result.objects) == 2
+    assert result.query_profile is not None
+    assert len(result.query_profile.shards) > 0
+
+    shard = result.query_profile.shards[0]
+    assert "vector" in shard.searches
+    vector_profile = shard.searches["vector"]
+    assert_common_profile(vector_profile)
+
+    assert "vector_search_took" in vector_profile.details
+    assert_go_duration(
+        vector_profile.details["vector_search_took"], "vector_search_took"
+    )
+
+    assert "hnsw_flat_search" in vector_profile.details
+    assert vector_profile.details["hnsw_flat_search"] in ("true", "false")
+
+    layer_keys = [
+        k for k in vector_profile.details if k.startswith("knn_search_layer_")
+    ]
+    assert len(layer_keys) > 0, "Expected at least one knn_search_layer_*_took key"
+    for k in layer_keys:
+        assert_go_duration(vector_profile.details[k], k)
+
+    assert "objects_took" in vector_profile.details
+    assert_go_duration(vector_profile.details["objects_took"], "objects_took")
+
+
+def test_bm25_with_query_profile(collection_with_data):
+    """Test that query profiling works with BM25 keyword search."""
+    result = collection_with_data.query.bm25(
+        query="hello",
+        return_metadata=MetadataQuery(query_profile=True, score=True),
+    )
+    assert result.query_profile is not None
+    assert len(result.query_profile.shards) > 0
+
+    shard = result.query_profile.shards[0]
+    assert "keyword" in shard.searches
+    keyword_profile = shard.searches["keyword"]
+    assert_common_profile(keyword_profile)
+
+    assert "kwd_method" in keyword_profile.details
+    assert keyword_profile.details["kwd_method"] != ""
+
+    assert "kwd_time" in keyword_profile.details
+    assert_go_duration(keyword_profile.details["kwd_time"], "kwd_time")
+
+    assert "kwd_1_tok_time" in keyword_profile.details
+    assert_go_duration(keyword_profile.details["kwd_1_tok_time"], "kwd_1_tok_time")
+
+    assert "kwd_6_res_count" in keyword_profile.details
+    assert keyword_profile.details["kwd_6_res_count"].isdigit()
+    assert int(keyword_profile.details["kwd_6_res_count"]) >= 0
+
+
+def test_hybrid_with_query_profile(collection_with_data):
+    """Test that query profiling works with hybrid search (both vector and keyword)."""
+    result = collection_with_data.query.hybrid(
+        query="hello",
+        vector=[1.0, 0.0, 0.0],
+        return_metadata=MetadataQuery(query_profile=True),
+        limit=2,
+    )
+    assert result.query_profile is not None
+    assert len(result.query_profile.shards) > 0
+
+    shard = result.query_profile.shards[0]
+    assert "vector" in shard.searches, "Hybrid should produce a 'vector' profile"
+    assert "keyword" in shard.searches, "Hybrid should produce a 'keyword' profile"
+
+    assert_common_profile(shard.searches["vector"])
+    assert "vector_search_took" in shard.searches["vector"].details
+
+    assert_common_profile(shard.searches["keyword"])
+    assert "kwd_method" in shard.searches["keyword"].details
+
+
+def test_near_vector_group_by_with_query_profile(collection_with_data):
+    """Test that query profiling works with group_by (mirrors C# QueryProfiling_NearText_GroupBy_Returns_Profile)."""
+    from weaviate.collections.classes.grpc import GroupBy
+
+    result = collection_with_data.query.near_vector(
+        near_vector=[1.0, 0.0, 0.0],
+        return_metadata=MetadataQuery(query_profile=True),
+        group_by=GroupBy(prop="text", objects_per_group=1, number_of_groups=3),
+    )
+    assert result.query_profile is not None
+    assert len(result.query_profile.shards) > 0
+
+    shard = result.query_profile.shards[0]
+    assert "vector" in shard.searches
+    assert_common_profile(shard.searches["vector"])
+
+
+def test_no_query_profile_when_not_requested(collection_with_data):
+    """Test that query_profile is None when not requested."""
+    result = collection_with_data.query.fetch_objects(
+        return_metadata=MetadataQuery(distance=True),
+    )
+    assert result.query_profile is None
+
+
+def test_query_profile_with_metadata_list(collection_with_data):
+    """Test that query profiling works when using list-style metadata."""
+    result = collection_with_data.query.near_vector(
+        near_vector=[1.0, 0.0, 0.0],
+        return_metadata=["query_profile", "distance"],
+        limit=2,
+    )
+    assert result.query_profile is not None
+    assert len(result.query_profile.shards) > 0
+
+    shard = result.query_profile.shards[0]
+    assert "vector" in shard.searches
+    assert_common_profile(shard.searches["vector"])
+
+
+def test_query_profile_details_are_strings(collection_with_data):
+    """Test that all detail keys and values are non-empty strings."""
+    result = collection_with_data.query.near_vector(
+        near_vector=[1.0, 0.0, 0.0],
+        return_metadata=MetadataQuery(query_profile=True),
+        limit=1,
+    )
+    assert result.query_profile is not None
+    for shard in result.query_profile.shards:
+        assert len(shard.searches) > 0, "Shard should have at least one search profile"
+        for search_type, profile in shard.searches.items():
+            assert isinstance(search_type, str) and search_type != ""
+            assert len(profile.details) > 0
+            for key, value in profile.details.items():
+                assert isinstance(key, str) and key != ""
+                assert isinstance(value, str) and value != ""
diff --git a/weaviate/collections/classes/grpc.py b/weaviate/collections/classes/grpc.py
index bff0e35ca..20ece3bc0 100644
--- a/weaviate/collections/classes/grpc.py
+++ b/weaviate/collections/classes/grpc.py
@@ -90,6 +90,7 @@ class MetadataQuery(_WeaviateInput):
     score: bool = Field(default=False)
     explain_score: bool = Field(default=False)
     is_consistent: bool = Field(default=False)
+    query_profile: bool = Field(default=False)
 
     @classmethod
     def full(cls) -> "MetadataQuery":
@@ -117,6 +118,7 @@ class _MetadataQuery:
     explain_score: bool = False
     is_consistent: bool = False
     vectors: Optional[List[str]] = None
+    query_profile: bool = False
 
     @classmethod
     def from_public(
@@ -138,6 +140,7 @@ def from_public(
                 score=public.score,
                 explain_score=public.explain_score,
                 is_consistent=public.is_consistent,
+                query_profile=public.query_profile,
             )
         )
 
@@ -152,6 +155,7 @@ def from_public(
             "score",
             "explain_score",
             "is_consistent",
+            "query_profile",
         ]
     ],
     MetadataQuery,
diff --git a/weaviate/collections/classes/internal.py b/weaviate/collections/classes/internal.py
index f4d542f66..55a69c5cd 100644
--- a/weaviate/collections/classes/internal.py
+++ b/weaviate/collections/classes/internal.py
@@ -90,6 +90,29 @@ def _is_empty(self) -> bool:
         )
 
 
+@dataclass
+class SearchProfileReturn:
+    """Profiling details for a single search type within a shard."""
+
+    details: Dict[str, str]
+
+
+@dataclass
+class ShardProfileReturn:
+    """Profiling data for a single shard."""
+
+    name: str
+    node: str
+    searches: Dict[str, SearchProfileReturn]
+
+
+@dataclass
+class QueryProfileReturn:
+    """Per-shard query profiling data returned when `query_profile=True` is set in metadata."""
+
+    shards: List[ShardProfileReturn]
+
+
 @dataclass
 class GroupByMetadataReturn:
     """Metadata of an object returned by a group by query."""
@@ -210,6 +233,7 @@ class GenerativeReturn(Generic[P, R]):
     __generated: Optional[str]
     objects: List[GenerativeObject[P, R]]
     generative: Optional[GenerativeGrouped]
+    query_profile: Optional[QueryProfileReturn]
 
     # init required because of nuances of dataclass when defining @property generated and private var __generated
     def __init__(
@@ -217,10 +241,12 @@ def __init__(
         generated: Optional[str],
         objects: List[GenerativeObject[P, R]],
         generative: Optional[GenerativeGrouped],
+        query_profile: Optional[QueryProfileReturn] = None,
     ) -> None:
         self.__generated = generated
         self.objects = objects
         self.generative = generative
+        self.query_profile = query_profile
 
     @property
     @deprecated(
@@ -257,6 +283,7 @@ class GenerativeGroupByReturn(Generic[P, R]):
     objects: List[GroupByObject[P, R]]
     groups: Dict[str, GenerativeGroup[P, R]]
     generated: Optional[str]
+    query_profile: Optional[QueryProfileReturn] = None
 
 
 @dataclass
@@ -265,6 +292,7 @@ class GroupByReturn(Generic[P, R]):
 
     objects: List[GroupByObject[P, R]]
     groups: Dict[str, Group[P, R]]
+    query_profile: Optional[QueryProfileReturn] = None
 
 
 @dataclass
@@ -272,6 +300,7 @@ class QueryReturn(Generic[P, R]):
     """The return type of a query within the `.query` namespace of a collection."""
 
     objects: List[Object[P, R]]
+    query_profile: Optional[QueryProfileReturn] = None
 
 
 _GQLEntryReturnType: TypeAlias = Dict[str, List[Dict[str, Any]]]
diff --git a/weaviate/collections/grpc/query.py b/weaviate/collections/grpc/query.py
index 1ac014ada..d635a3f2f 100644
--- a/weaviate/collections/grpc/query.py
+++ b/weaviate/collections/grpc/query.py
@@ -521,6 +521,7 @@ def _metadata_to_grpc(self, metadata: _MetadataQuery) -> search_get_pb2.Metadata
             score=metadata.score,
             is_consistent=metadata.is_consistent,
             vectors=metadata.vectors,
+            query_profile=metadata.query_profile,
         )
 
     def __resolve_property(self, prop: QueryNested) -> search_get_pb2.ObjectPropertiesRequest:
diff --git a/weaviate/collections/queries/base_executor.py b/weaviate/collections/queries/base_executor.py
index e721d9edb..f3eb36399 100644
--- a/weaviate/collections/queries/base_executor.py
+++ b/weaviate/collections/queries/base_executor.py
@@ -40,9 +40,12 @@
     GroupByReturn,
     MetadataReturn,
     Object,
+    QueryProfileReturn,
     QueryReturn,
     ReturnProperties,
     ReturnReferences,
+    SearchProfileReturn,
+    ShardProfileReturn,
     WeaviateProperties,
     _CrossReference,
     _extract_properties_from_data_model,
@@ -53,7 +56,7 @@
 from weaviate.collections.grpc.query import _QueryGRPC
 from weaviate.collections.grpc.shared import _ByteOps, _Unpack
 from weaviate.connect.v4 import ConnectionType
-from weaviate.exceptions import WeaviateInvalidInputError
+from weaviate.exceptions import WeaviateInvalidInputError, WeaviateUnsupportedFeatureError
 from weaviate.proto.v1 import base_pb2, generative_pb2, properties_pb2, search_get_pb2
 from weaviate.types import INCLUDE_VECTOR
 from weaviate.util import (
@@ -452,6 +455,25 @@ def __result_to_group_by_object(
             belongs_to_group=group_name,
         )
 
+    def __extract_query_profile(
+        self, res: search_get_pb2.SearchReply
+    ) -> Optional[QueryProfileReturn]:
+        if not res.HasField("query_profile"):
+            return None
+        return QueryProfileReturn(
+            shards=[
+                ShardProfileReturn(
+                    name=shard.name,
+                    node=shard.node,
+                    searches={
+                        key: SearchProfileReturn(details=dict(profile.details))
+                        for key, profile in shard.searches.items()
+                    },
+                )
+                for shard in res.query_profile.shards
+            ]
+        )
+
     def _result_to_query_return(
         self,
         res: search_get_pb2.SearchReply,
@@ -461,7 +483,8 @@ def _result_to_query_return(
             objects=[
                 self.__result_to_query_object(obj.properties, obj.metadata, options)
                 for obj in res.results
-            ]
+            ],
+            query_profile=self.__extract_query_profile(res),
         )
 
     def _result_to_generative_query_return(
@@ -480,6 +503,7 @@ def _result_to_generative_query_return(
             generative=self.__extract_generative_grouped_from_generative(
                 res.generative_grouped_results
             ),
+            query_profile=self.__extract_query_profile(res),
         )
 
     def _result_to_generative_return(
@@ -507,7 +531,11 @@ def _result_to_groupby_return(
         objects_group_by: List[GroupByObject] = [
             obj for group in groups.values() for obj in group.objects
         ]
-        return GroupByReturn(objects=objects_group_by, groups=groups)
+        return GroupByReturn(
+            objects=objects_group_by,
+            groups=groups,
+            query_profile=self.__extract_query_profile(res),
+        )
 
     def _result_to_generative_groupby_return(
         self,
@@ -537,6 +565,7 @@ def _result_to_generative_groupby_return(
             generated=(
                 res.generative_grouped_result if res.generative_grouped_result != "" else None
             ),
+            query_profile=self.__extract_query_profile(res),
         )
 
     def _result_to_query_or_groupby_return(
@@ -615,6 +644,13 @@ def _parse_return_metadata(
             ret_md = cast(MetadataQuery, return_metadata)
         else:
             ret_md = MetadataQuery(**{str(prop): True for prop in return_metadata})
+
+        if ret_md is not None and ret_md.query_profile:
+            if self._connection._weaviate_version.is_lower_than(1, 36, 9):
+                raise WeaviateUnsupportedFeatureError(
+                    "Query profiling", str(self._connection._weaviate_version), "1.36.9"
+                )
+
         return _MetadataQuery.from_public(ret_md, include_vector)
 
     def _parse_return_references(
diff --git a/weaviate/outputs/query.py b/weaviate/outputs/query.py
index 42ede14b3..625d669d8 100644
--- a/weaviate/outputs/query.py
+++ b/weaviate/outputs/query.py
@@ -32,11 +32,14 @@
     Object,
     ObjectSingleReturn,
     QueryNearMediaReturnType,
+    QueryProfileReturn,
     QueryReturn,
     QueryReturnType,
     QuerySingleReturn,
     ReferenceInput,
     ReferenceInputs,
+    SearchProfileReturn,
+    ShardProfileReturn,
 )
 from weaviate.collections.classes.types import (
     GeoCoordinate,
@@ -75,11 +78,14 @@
     "GenerativeGroup",
     "PhoneNumberType",
     "QueryNearMediaReturnType",
+    "QueryProfileReturn",
     "QueryReturnType",
     "QueryReturn",
     "QuerySingleReturn",
     "ReferenceInput",
     "ReferenceInputs",
+    "SearchProfileReturn",
+    "ShardProfileReturn",
     "Sorting",
     "TargetVectorJoinType",
     "WeaviateField",
diff --git a/weaviate/proto/v1/v4216/v1/search_get_pb2.py b/weaviate/proto/v1/v4216/v1/search_get_pb2.py
index 7885e57d6..78f20076e 100644
--- a/weaviate/proto/v1/v4216/v1/search_get_pb2.py
+++ b/weaviate/proto/v1/v4216/v1/search_get_pb2.py
@@ -17,7 +17,7 @@
 from weaviate.proto.v1.v4216.v1 import properties_pb2 as v1_dot_properties__pb2
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x13v1/search_get.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\x1a\x14v1/base_search.proto\x1a\x13v1/generative.proto\x1a\x13v1/properties.proto\"\x9c\x0b\n\rSearchRequest\x12\x12\n\ncollection\x18\x01 \x01(\t\x12\x0e\n\x06tenant\x18\n \x01(\t\x12=\n\x11\x63onsistency_level\x18\x0b \x01(\x0e\x32\x1d.weaviate.v1.ConsistencyLevelH\x00\x88\x01\x01\x12\x37\n\nproperties\x18\x14 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequestH\x01\x88\x01\x01\x12\x33\n\x08metadata\x18\x15 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequestH\x02\x88\x01\x01\x12+\n\x08group_by\x18\x16 \x01(\x0b\x32\x14.weaviate.v1.GroupByH\x03\x88\x01\x01\x12\r\n\x05limit\x18\x1e \x01(\r\x12\x0e\n\x06offset\x18\x1f \x01(\r\x12\x0f\n\x07\x61utocut\x18  \x01(\r\x12\r\n\x05\x61\x66ter\x18! \x01(\t\x12$\n\x07sort_by\x18\" \x03(\x0b\x32\x13.weaviate.v1.SortBy\x12*\n\x07\x66ilters\x18( \x01(\x0b\x32\x14.weaviate.v1.FiltersH\x04\x88\x01\x01\x12/\n\rhybrid_search\x18) \x01(\x0b\x32\x13.weaviate.v1.HybridH\x05\x88\x01\x01\x12+\n\x0b\x62m25_search\x18* \x01(\x0b\x32\x11.weaviate.v1.BM25H\x06\x88\x01\x01\x12\x31\n\x0bnear_vector\x18+ \x01(\x0b\x32\x17.weaviate.v1.NearVectorH\x07\x88\x01\x01\x12\x31\n\x0bnear_object\x18, \x01(\x0b\x32\x17.weaviate.v1.NearObjectH\x08\x88\x01\x01\x12\x33\n\tnear_text\x18- \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearchH\t\x88\x01\x01\x12\x35\n\nnear_image\x18. \x01(\x0b\x32\x1c.weaviate.v1.NearImageSearchH\n\x88\x01\x01\x12\x35\n\nnear_audio\x18/ \x01(\x0b\x32\x1c.weaviate.v1.NearAudioSearchH\x0b\x88\x01\x01\x12\x35\n\nnear_video\x18\x30 \x01(\x0b\x32\x1c.weaviate.v1.NearVideoSearchH\x0c\x88\x01\x01\x12\x35\n\nnear_depth\x18\x31 \x01(\x0b\x32\x1c.weaviate.v1.NearDepthSearchH\r\x88\x01\x01\x12\x39\n\x0cnear_thermal\x18\x32 \x01(\x0b\x32\x1e.weaviate.v1.NearThermalSearchH\x0e\x88\x01\x01\x12\x31\n\x08near_imu\x18\x33 \x01(\x0b\x32\x1a.weaviate.v1.NearIMUSearchH\x0f\x88\x01\x01\x12\x36\n\ngenerative\x18< \x01(\x0b\x32\x1d.weaviate.v1.GenerativeSearchH\x10\x88\x01\x01\x12(\n\x06rerank\x18= \x01(\x0b\x32\x13.weaviate.v1.RerankH\x11\x88\x01\x01\x12\x18\n\x0cuses_123_api\x18\x64 \x01(\x08\x42\x02\x18\x01\x12\x18\n\x0cuses_125_api\x18\x65 \x01(\x08\x42\x02\x18\x01\x12\x14\n\x0cuses_127_api\x18\x66 \x01(\x08\x42\x14\n\x12_consistency_levelB\r\n\x0b_propertiesB\x0b\n\t_metadataB\x0b\n\t_group_byB\n\n\x08_filtersB\x10\n\x0e_hybrid_searchB\x0e\n\x0c_bm25_searchB\x0e\n\x0c_near_vectorB\x0e\n\x0c_near_objectB\x0c\n\n_near_textB\r\n\x0b_near_imageB\r\n\x0b_near_audioB\r\n\x0b_near_videoB\r\n\x0b_near_depthB\x0f\n\r_near_thermalB\x0b\n\t_near_imuB\r\n\x0b_generativeB\t\n\x07_rerank\"L\n\x07GroupBy\x12\x0c\n\x04path\x18\x01 \x03(\t\x12\x18\n\x10number_of_groups\x18\x02 \x01(\x05\x12\x19\n\x11objects_per_group\x18\x03 \x01(\x05\")\n\x06SortBy\x12\x11\n\tascending\x18\x01 \x01(\x08\x12\x0c\n\x04path\x18\x02 \x03(\t\"\xdd\x01\n\x0fMetadataRequest\x12\x0c\n\x04uuid\x18\x01 \x01(\x08\x12\x0e\n\x06vector\x18\x02 \x01(\x08\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x04 \x01(\x08\x12\x10\n\x08\x64istance\x18\x05 \x01(\x08\x12\x11\n\tcertainty\x18\x06 \x01(\x08\x12\r\n\x05score\x18\x07 \x01(\x08\x12\x15\n\rexplain_score\x18\x08 \x01(\x08\x12\x15\n\ris_consistent\x18\t \x01(\x08\x12\x0f\n\x07vectors\x18\n \x03(\t\"\xd1\x01\n\x11PropertiesRequest\x12\x1a\n\x12non_ref_properties\x18\x01 \x03(\t\x12\x39\n\x0eref_properties\x18\x02 \x03(\x0b\x32!.weaviate.v1.RefPropertiesRequest\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\x12$\n\x1creturn_all_nonref_properties\x18\x0b \x01(\x08\"\x8b\x01\n\x17ObjectPropertiesRequest\x12\x11\n\tprop_name\x18\x01 \x01(\t\x12\x1c\n\x14primitive_properties\x18\x02 \x03(\t\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\"\xb1\x01\n\x14RefPropertiesRequest\x12\x1a\n\x12reference_property\x18\x01 \x01(\t\x12\x32\n\nproperties\x18\x02 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequest\x12.\n\x08metadata\x18\x03 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequest\x12\x19\n\x11target_collection\x18\x04 \x01(\t\"8\n\x06Rerank\x12\x10\n\x08property\x18\x01 \x01(\t\x12\x12\n\x05query\x18\x02 \x01(\tH\x00\x88\x01\x01\x42\x08\n\x06_query\"\xae\x02\n\x0bSearchReply\x12\x0c\n\x04took\x18\x01 \x01(\x02\x12*\n\x07results\x18\x02 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12*\n\x19generative_grouped_result\x18\x03 \x01(\tB\x02\x18\x01H\x00\x88\x01\x01\x12\x34\n\x10group_by_results\x18\x04 \x03(\x0b\x32\x1a.weaviate.v1.GroupByResult\x12\x46\n\x1agenerative_grouped_results\x18\x05 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x01\x88\x01\x01\x42\x1c\n\x1a_generative_grouped_resultB\x1d\n\x1b_generative_grouped_results\"\x1c\n\x0bRerankReply\x12\r\n\x05score\x18\x01 \x01(\x01\"\xe9\x02\n\rGroupByResult\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x14\n\x0cmin_distance\x18\x02 \x01(\x02\x12\x14\n\x0cmax_distance\x18\x03 \x01(\x02\x12\x19\n\x11number_of_objects\x18\x04 \x01(\x03\x12*\n\x07objects\x18\x05 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12-\n\x06rerank\x18\x06 \x01(\x0b\x32\x18.weaviate.v1.RerankReplyH\x00\x88\x01\x01\x12\x39\n\ngenerative\x18\x07 \x01(\x0b\x32\x1c.weaviate.v1.GenerativeReplyB\x02\x18\x01H\x01\x88\x01\x01\x12=\n\x11generative_result\x18\x08 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x02\x88\x01\x01\x42\t\n\x07_rerankB\r\n\x0b_generativeB\x14\n\x12_generative_result\"\xb7\x01\n\x0cSearchResult\x12\x31\n\nproperties\x18\x01 \x01(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12-\n\x08metadata\x18\x02 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12\x36\n\ngenerative\x18\x03 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x00\x88\x01\x01\x42\r\n\x0b_generative\"\xf7\x04\n\x0eMetadataResult\x12\n\n\x02id\x18\x01 \x01(\t\x12\x12\n\x06vector\x18\x02 \x03(\x02\x42\x02\x18\x01\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x03\x12\"\n\x1a\x63reation_time_unix_present\x18\x04 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x05 \x01(\x03\x12%\n\x1dlast_update_time_unix_present\x18\x06 \x01(\x08\x12\x10\n\x08\x64istance\x18\x07 \x01(\x02\x12\x18\n\x10\x64istance_present\x18\x08 \x01(\x08\x12\x11\n\tcertainty\x18\t \x01(\x02\x12\x19\n\x11\x63\x65rtainty_present\x18\n \x01(\x08\x12\r\n\x05score\x18\x0b \x01(\x02\x12\x15\n\rscore_present\x18\x0c \x01(\x08\x12\x15\n\rexplain_score\x18\r \x01(\t\x12\x1d\n\x15\x65xplain_score_present\x18\x0e \x01(\x08\x12\x1a\n\ris_consistent\x18\x0f \x01(\x08H\x00\x88\x01\x01\x12\x16\n\ngenerative\x18\x10 \x01(\tB\x02\x18\x01\x12\x1e\n\x12generative_present\x18\x11 \x01(\x08\x42\x02\x18\x01\x12\x1d\n\x15is_consistent_present\x18\x12 \x01(\x08\x12\x14\n\x0cvector_bytes\x18\x13 \x01(\x0c\x12\x13\n\x0bid_as_bytes\x18\x14 \x01(\x0c\x12\x14\n\x0crerank_score\x18\x15 \x01(\x01\x12\x1c\n\x14rerank_score_present\x18\x16 \x01(\x08\x12%\n\x07vectors\x18\x17 \x03(\x0b\x32\x14.weaviate.v1.VectorsB\x10\n\x0e_is_consistent\"\x88\x02\n\x10PropertiesResult\x12\x33\n\tref_props\x18\x02 \x03(\x0b\x32 .weaviate.v1.RefPropertiesResult\x12\x19\n\x11target_collection\x18\x03 \x01(\t\x12-\n\x08metadata\x18\x04 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12.\n\rnon_ref_props\x18\x0b \x01(\x0b\x32\x17.weaviate.v1.Properties\x12\x1b\n\x13ref_props_requested\x18\x0c \x01(\x08J\x04\x08\x01\x10\x02J\x04\x08\x05\x10\x06J\x04\x08\x06\x10\x07J\x04\x08\x07\x10\x08J\x04\x08\x08\x10\tJ\x04\x08\t\x10\nJ\x04\x08\n\x10\x0b\"[\n\x13RefPropertiesResult\x12\x31\n\nproperties\x18\x01 \x03(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12\x11\n\tprop_name\x18\x02 \x01(\tBs\n#io.weaviate.client.grpc.protocol.v1B\x16WeaviateProtoSearchGetZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x13v1/search_get.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\x1a\x14v1/base_search.proto\x1a\x13v1/generative.proto\x1a\x13v1/properties.proto\"\x9c\x0b\n\rSearchRequest\x12\x12\n\ncollection\x18\x01 \x01(\t\x12\x0e\n\x06tenant\x18\n \x01(\t\x12=\n\x11\x63onsistency_level\x18\x0b \x01(\x0e\x32\x1d.weaviate.v1.ConsistencyLevelH\x00\x88\x01\x01\x12\x37\n\nproperties\x18\x14 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequestH\x01\x88\x01\x01\x12\x33\n\x08metadata\x18\x15 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequestH\x02\x88\x01\x01\x12+\n\x08group_by\x18\x16 \x01(\x0b\x32\x14.weaviate.v1.GroupByH\x03\x88\x01\x01\x12\r\n\x05limit\x18\x1e \x01(\r\x12\x0e\n\x06offset\x18\x1f \x01(\r\x12\x0f\n\x07\x61utocut\x18  \x01(\r\x12\r\n\x05\x61\x66ter\x18! \x01(\t\x12$\n\x07sort_by\x18\" \x03(\x0b\x32\x13.weaviate.v1.SortBy\x12*\n\x07\x66ilters\x18( \x01(\x0b\x32\x14.weaviate.v1.FiltersH\x04\x88\x01\x01\x12/\n\rhybrid_search\x18) \x01(\x0b\x32\x13.weaviate.v1.HybridH\x05\x88\x01\x01\x12+\n\x0b\x62m25_search\x18* \x01(\x0b\x32\x11.weaviate.v1.BM25H\x06\x88\x01\x01\x12\x31\n\x0bnear_vector\x18+ \x01(\x0b\x32\x17.weaviate.v1.NearVectorH\x07\x88\x01\x01\x12\x31\n\x0bnear_object\x18, \x01(\x0b\x32\x17.weaviate.v1.NearObjectH\x08\x88\x01\x01\x12\x33\n\tnear_text\x18- \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearchH\t\x88\x01\x01\x12\x35\n\nnear_image\x18. \x01(\x0b\x32\x1c.weaviate.v1.NearImageSearchH\n\x88\x01\x01\x12\x35\n\nnear_audio\x18/ \x01(\x0b\x32\x1c.weaviate.v1.NearAudioSearchH\x0b\x88\x01\x01\x12\x35\n\nnear_video\x18\x30 \x01(\x0b\x32\x1c.weaviate.v1.NearVideoSearchH\x0c\x88\x01\x01\x12\x35\n\nnear_depth\x18\x31 \x01(\x0b\x32\x1c.weaviate.v1.NearDepthSearchH\r\x88\x01\x01\x12\x39\n\x0cnear_thermal\x18\x32 \x01(\x0b\x32\x1e.weaviate.v1.NearThermalSearchH\x0e\x88\x01\x01\x12\x31\n\x08near_imu\x18\x33 \x01(\x0b\x32\x1a.weaviate.v1.NearIMUSearchH\x0f\x88\x01\x01\x12\x36\n\ngenerative\x18< \x01(\x0b\x32\x1d.weaviate.v1.GenerativeSearchH\x10\x88\x01\x01\x12(\n\x06rerank\x18= \x01(\x0b\x32\x13.weaviate.v1.RerankH\x11\x88\x01\x01\x12\x18\n\x0cuses_123_api\x18\x64 \x01(\x08\x42\x02\x18\x01\x12\x18\n\x0cuses_125_api\x18\x65 \x01(\x08\x42\x02\x18\x01\x12\x14\n\x0cuses_127_api\x18\x66 \x01(\x08\x42\x14\n\x12_consistency_levelB\r\n\x0b_propertiesB\x0b\n\t_metadataB\x0b\n\t_group_byB\n\n\x08_filtersB\x10\n\x0e_hybrid_searchB\x0e\n\x0c_bm25_searchB\x0e\n\x0c_near_vectorB\x0e\n\x0c_near_objectB\x0c\n\n_near_textB\r\n\x0b_near_imageB\r\n\x0b_near_audioB\r\n\x0b_near_videoB\r\n\x0b_near_depthB\x0f\n\r_near_thermalB\x0b\n\t_near_imuB\r\n\x0b_generativeB\t\n\x07_rerank\"L\n\x07GroupBy\x12\x0c\n\x04path\x18\x01 \x03(\t\x12\x18\n\x10number_of_groups\x18\x02 \x01(\x05\x12\x19\n\x11objects_per_group\x18\x03 \x01(\x05\")\n\x06SortBy\x12\x11\n\tascending\x18\x01 \x01(\x08\x12\x0c\n\x04path\x18\x02 \x03(\t\"\xf4\x01\n\x0fMetadataRequest\x12\x0c\n\x04uuid\x18\x01 \x01(\x08\x12\x0e\n\x06vector\x18\x02 \x01(\x08\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x04 \x01(\x08\x12\x10\n\x08\x64istance\x18\x05 \x01(\x08\x12\x11\n\tcertainty\x18\x06 \x01(\x08\x12\r\n\x05score\x18\x07 \x01(\x08\x12\x15\n\rexplain_score\x18\x08 \x01(\x08\x12\x15\n\ris_consistent\x18\t \x01(\x08\x12\x0f\n\x07vectors\x18\n \x03(\t\x12\x15\n\rquery_profile\x18\x0b \x01(\x08\"\xd1\x01\n\x11PropertiesRequest\x12\x1a\n\x12non_ref_properties\x18\x01 \x03(\t\x12\x39\n\x0eref_properties\x18\x02 \x03(\x0b\x32!.weaviate.v1.RefPropertiesRequest\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\x12$\n\x1creturn_all_nonref_properties\x18\x0b \x01(\x08\"\x8b\x01\n\x17ObjectPropertiesRequest\x12\x11\n\tprop_name\x18\x01 \x01(\t\x12\x1c\n\x14primitive_properties\x18\x02 \x03(\t\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\"\xb1\x01\n\x14RefPropertiesRequest\x12\x1a\n\x12reference_property\x18\x01 \x01(\t\x12\x32\n\nproperties\x18\x02 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequest\x12.\n\x08metadata\x18\x03 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequest\x12\x19\n\x11target_collection\x18\x04 \x01(\t\"8\n\x06Rerank\x12\x10\n\x08property\x18\x01 \x01(\t\x12\x12\n\x05query\x18\x02 \x01(\tH\x00\x88\x01\x01\x42\x08\n\x06_query\"\xf7\x02\n\x0bSearchReply\x12\x0c\n\x04took\x18\x01 \x01(\x02\x12*\n\x07results\x18\x02 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12*\n\x19generative_grouped_result\x18\x03 \x01(\tB\x02\x18\x01H\x00\x88\x01\x01\x12\x34\n\x10group_by_results\x18\x04 \x03(\x0b\x32\x1a.weaviate.v1.GroupByResult\x12\x46\n\x1agenerative_grouped_results\x18\x05 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x01\x88\x01\x01\x12\x35\n\rquery_profile\x18\x06 \x01(\x0b\x32\x19.weaviate.v1.QueryProfileH\x02\x88\x01\x01\x42\x1c\n\x1a_generative_grouped_resultB\x1d\n\x1b_generative_grouped_resultsB\x10\n\x0e_query_profile\"\x9e\x03\n\x0cQueryProfile\x12\x36\n\x06shards\x18\x01 \x03(\x0b\x32&.weaviate.v1.QueryProfile.ShardProfile\x1a\x86\x01\n\rSearchProfile\x12\x45\n\x07\x64\x65tails\x18\x01 \x03(\x0b\x32\x34.weaviate.v1.QueryProfile.SearchProfile.DetailsEntry\x1a.\n\x0c\x44\x65tailsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01\x1a\xcc\x01\n\x0cShardProfile\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x0c\n\x04node\x18\x02 \x01(\t\x12\x46\n\x08searches\x18\x03 \x03(\x0b\x32\x34.weaviate.v1.QueryProfile.ShardProfile.SearchesEntry\x1aX\n\rSearchesEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\x36\n\x05value\x18\x02 \x01(\x0b\x32\'.weaviate.v1.QueryProfile.SearchProfile:\x02\x38\x01\"\x1c\n\x0bRerankReply\x12\r\n\x05score\x18\x01 \x01(\x01\"\xe9\x02\n\rGroupByResult\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x14\n\x0cmin_distance\x18\x02 \x01(\x02\x12\x14\n\x0cmax_distance\x18\x03 \x01(\x02\x12\x19\n\x11number_of_objects\x18\x04 \x01(\x03\x12*\n\x07objects\x18\x05 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12-\n\x06rerank\x18\x06 \x01(\x0b\x32\x18.weaviate.v1.RerankReplyH\x00\x88\x01\x01\x12\x39\n\ngenerative\x18\x07 \x01(\x0b\x32\x1c.weaviate.v1.GenerativeReplyB\x02\x18\x01H\x01\x88\x01\x01\x12=\n\x11generative_result\x18\x08 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x02\x88\x01\x01\x42\t\n\x07_rerankB\r\n\x0b_generativeB\x14\n\x12_generative_result\"\xb7\x01\n\x0cSearchResult\x12\x31\n\nproperties\x18\x01 \x01(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12-\n\x08metadata\x18\x02 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12\x36\n\ngenerative\x18\x03 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x00\x88\x01\x01\x42\r\n\x0b_generative\"\xf7\x04\n\x0eMetadataResult\x12\n\n\x02id\x18\x01 \x01(\t\x12\x12\n\x06vector\x18\x02 \x03(\x02\x42\x02\x18\x01\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x03\x12\"\n\x1a\x63reation_time_unix_present\x18\x04 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x05 \x01(\x03\x12%\n\x1dlast_update_time_unix_present\x18\x06 \x01(\x08\x12\x10\n\x08\x64istance\x18\x07 \x01(\x02\x12\x18\n\x10\x64istance_present\x18\x08 \x01(\x08\x12\x11\n\tcertainty\x18\t \x01(\x02\x12\x19\n\x11\x63\x65rtainty_present\x18\n \x01(\x08\x12\r\n\x05score\x18\x0b \x01(\x02\x12\x15\n\rscore_present\x18\x0c \x01(\x08\x12\x15\n\rexplain_score\x18\r \x01(\t\x12\x1d\n\x15\x65xplain_score_present\x18\x0e \x01(\x08\x12\x1a\n\ris_consistent\x18\x0f \x01(\x08H\x00\x88\x01\x01\x12\x16\n\ngenerative\x18\x10 \x01(\tB\x02\x18\x01\x12\x1e\n\x12generative_present\x18\x11 \x01(\x08\x42\x02\x18\x01\x12\x1d\n\x15is_consistent_present\x18\x12 \x01(\x08\x12\x14\n\x0cvector_bytes\x18\x13 \x01(\x0c\x12\x13\n\x0bid_as_bytes\x18\x14 \x01(\x0c\x12\x14\n\x0crerank_score\x18\x15 \x01(\x01\x12\x1c\n\x14rerank_score_present\x18\x16 \x01(\x08\x12%\n\x07vectors\x18\x17 \x03(\x0b\x32\x14.weaviate.v1.VectorsB\x10\n\x0e_is_consistent\"\x88\x02\n\x10PropertiesResult\x12\x33\n\tref_props\x18\x02 \x03(\x0b\x32 .weaviate.v1.RefPropertiesResult\x12\x19\n\x11target_collection\x18\x03 \x01(\t\x12-\n\x08metadata\x18\x04 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12.\n\rnon_ref_props\x18\x0b \x01(\x0b\x32\x17.weaviate.v1.Properties\x12\x1b\n\x13ref_props_requested\x18\x0c \x01(\x08J\x04\x08\x01\x10\x02J\x04\x08\x05\x10\x06J\x04\x08\x06\x10\x07J\x04\x08\x07\x10\x08J\x04\x08\x08\x10\tJ\x04\x08\t\x10\nJ\x04\x08\n\x10\x0b\"[\n\x13RefPropertiesResult\x12\x31\n\nproperties\x18\x01 \x03(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12\x11\n\tprop_name\x18\x02 \x01(\tBs\n#io.weaviate.client.grpc.protocol.v1B\x16WeaviateProtoSearchGetZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -31,6 +31,10 @@
   _SEARCHREQUEST.fields_by_name['uses_125_api']._serialized_options = b'\030\001'
   _SEARCHREPLY.fields_by_name['generative_grouped_result']._options = None
   _SEARCHREPLY.fields_by_name['generative_grouped_result']._serialized_options = b'\030\001'
+  _QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY._options = None
+  _QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY._serialized_options = b'8\001'
+  _QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY._options = None
+  _QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY._serialized_options = b'8\001'
   _GROUPBYRESULT.fields_by_name['generative']._options = None
   _GROUPBYRESULT.fields_by_name['generative']._serialized_options = b'\030\001'
   _METADATARESULT.fields_by_name['vector']._options = None
@@ -46,27 +50,37 @@
   _globals['_SORTBY']._serialized_start=1632
   _globals['_SORTBY']._serialized_end=1673
   _globals['_METADATAREQUEST']._serialized_start=1676
-  _globals['_METADATAREQUEST']._serialized_end=1897
-  _globals['_PROPERTIESREQUEST']._serialized_start=1900
-  _globals['_PROPERTIESREQUEST']._serialized_end=2109
-  _globals['_OBJECTPROPERTIESREQUEST']._serialized_start=2112
-  _globals['_OBJECTPROPERTIESREQUEST']._serialized_end=2251
-  _globals['_REFPROPERTIESREQUEST']._serialized_start=2254
-  _globals['_REFPROPERTIESREQUEST']._serialized_end=2431
-  _globals['_RERANK']._serialized_start=2433
-  _globals['_RERANK']._serialized_end=2489
-  _globals['_SEARCHREPLY']._serialized_start=2492
-  _globals['_SEARCHREPLY']._serialized_end=2794
-  _globals['_RERANKREPLY']._serialized_start=2796
-  _globals['_RERANKREPLY']._serialized_end=2824
-  _globals['_GROUPBYRESULT']._serialized_start=2827
-  _globals['_GROUPBYRESULT']._serialized_end=3188
-  _globals['_SEARCHRESULT']._serialized_start=3191
-  _globals['_SEARCHRESULT']._serialized_end=3374
-  _globals['_METADATARESULT']._serialized_start=3377
-  _globals['_METADATARESULT']._serialized_end=4008
-  _globals['_PROPERTIESRESULT']._serialized_start=4011
-  _globals['_PROPERTIESRESULT']._serialized_end=4275
-  _globals['_REFPROPERTIESRESULT']._serialized_start=4277
-  _globals['_REFPROPERTIESRESULT']._serialized_end=4368
+  _globals['_METADATAREQUEST']._serialized_end=1920
+  _globals['_PROPERTIESREQUEST']._serialized_start=1923
+  _globals['_PROPERTIESREQUEST']._serialized_end=2132
+  _globals['_OBJECTPROPERTIESREQUEST']._serialized_start=2135
+  _globals['_OBJECTPROPERTIESREQUEST']._serialized_end=2274
+  _globals['_REFPROPERTIESREQUEST']._serialized_start=2277
+  _globals['_REFPROPERTIESREQUEST']._serialized_end=2454
+  _globals['_RERANK']._serialized_start=2456
+  _globals['_RERANK']._serialized_end=2512
+  _globals['_SEARCHREPLY']._serialized_start=2515
+  _globals['_SEARCHREPLY']._serialized_end=2890
+  _globals['_QUERYPROFILE']._serialized_start=2893
+  _globals['_QUERYPROFILE']._serialized_end=3307
+  _globals['_QUERYPROFILE_SEARCHPROFILE']._serialized_start=2966
+  _globals['_QUERYPROFILE_SEARCHPROFILE']._serialized_end=3100
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._serialized_start=3054
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._serialized_end=3100
+  _globals['_QUERYPROFILE_SHARDPROFILE']._serialized_start=3103
+  _globals['_QUERYPROFILE_SHARDPROFILE']._serialized_end=3307
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._serialized_start=3219
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._serialized_end=3307
+  _globals['_RERANKREPLY']._serialized_start=3309
+  _globals['_RERANKREPLY']._serialized_end=3337
+  _globals['_GROUPBYRESULT']._serialized_start=3340
+  _globals['_GROUPBYRESULT']._serialized_end=3701
+  _globals['_SEARCHRESULT']._serialized_start=3704
+  _globals['_SEARCHRESULT']._serialized_end=3887
+  _globals['_METADATARESULT']._serialized_start=3890
+  _globals['_METADATARESULT']._serialized_end=4521
+  _globals['_PROPERTIESRESULT']._serialized_start=4524
+  _globals['_PROPERTIESRESULT']._serialized_end=4788
+  _globals['_REFPROPERTIESRESULT']._serialized_start=4790
+  _globals['_REFPROPERTIESRESULT']._serialized_end=4881
 # @@protoc_insertion_point(module_scope)
diff --git a/weaviate/proto/v1/v4216/v1/search_get_pb2.pyi b/weaviate/proto/v1/v4216/v1/search_get_pb2.pyi
index 9dd1ee0d0..a0300e672 100644
--- a/weaviate/proto/v1/v4216/v1/search_get_pb2.pyi
+++ b/weaviate/proto/v1/v4216/v1/search_get_pb2.pyi
@@ -88,7 +88,7 @@ class SortBy(_message.Message):
     def __init__(self, ascending: bool = ..., path: _Optional[_Iterable[str]] = ...) -> None: ...
 
 class MetadataRequest(_message.Message):
-    __slots__ = ["uuid", "vector", "creation_time_unix", "last_update_time_unix", "distance", "certainty", "score", "explain_score", "is_consistent", "vectors"]
+    __slots__ = ["uuid", "vector", "creation_time_unix", "last_update_time_unix", "distance", "certainty", "score", "explain_score", "is_consistent", "vectors", "query_profile"]
     UUID_FIELD_NUMBER: _ClassVar[int]
     VECTOR_FIELD_NUMBER: _ClassVar[int]
     CREATION_TIME_UNIX_FIELD_NUMBER: _ClassVar[int]
@@ -99,6 +99,7 @@ class MetadataRequest(_message.Message):
     EXPLAIN_SCORE_FIELD_NUMBER: _ClassVar[int]
     IS_CONSISTENT_FIELD_NUMBER: _ClassVar[int]
     VECTORS_FIELD_NUMBER: _ClassVar[int]
+    QUERY_PROFILE_FIELD_NUMBER: _ClassVar[int]
     uuid: bool
     vector: bool
     creation_time_unix: bool
@@ -109,7 +110,8 @@ class MetadataRequest(_message.Message):
     explain_score: bool
     is_consistent: bool
     vectors: _containers.RepeatedScalarFieldContainer[str]
-    def __init__(self, uuid: bool = ..., vector: bool = ..., creation_time_unix: bool = ..., last_update_time_unix: bool = ..., distance: bool = ..., certainty: bool = ..., score: bool = ..., explain_score: bool = ..., is_consistent: bool = ..., vectors: _Optional[_Iterable[str]] = ...) -> None: ...
+    query_profile: bool
+    def __init__(self, uuid: bool = ..., vector: bool = ..., creation_time_unix: bool = ..., last_update_time_unix: bool = ..., distance: bool = ..., certainty: bool = ..., score: bool = ..., explain_score: bool = ..., is_consistent: bool = ..., vectors: _Optional[_Iterable[str]] = ..., query_profile: bool = ...) -> None: ...
 
 class PropertiesRequest(_message.Message):
     __slots__ = ["non_ref_properties", "ref_properties", "object_properties", "return_all_nonref_properties"]
@@ -154,18 +156,54 @@ class Rerank(_message.Message):
     def __init__(self, property: _Optional[str] = ..., query: _Optional[str] = ...) -> None: ...
 
 class SearchReply(_message.Message):
-    __slots__ = ["took", "results", "generative_grouped_result", "group_by_results", "generative_grouped_results"]
+    __slots__ = ["took", "results", "generative_grouped_result", "group_by_results", "generative_grouped_results", "query_profile"]
     TOOK_FIELD_NUMBER: _ClassVar[int]
     RESULTS_FIELD_NUMBER: _ClassVar[int]
     GENERATIVE_GROUPED_RESULT_FIELD_NUMBER: _ClassVar[int]
     GROUP_BY_RESULTS_FIELD_NUMBER: _ClassVar[int]
     GENERATIVE_GROUPED_RESULTS_FIELD_NUMBER: _ClassVar[int]
+    QUERY_PROFILE_FIELD_NUMBER: _ClassVar[int]
     took: float
     results: _containers.RepeatedCompositeFieldContainer[SearchResult]
     generative_grouped_result: str
     group_by_results: _containers.RepeatedCompositeFieldContainer[GroupByResult]
     generative_grouped_results: _generative_pb2.GenerativeResult
-    def __init__(self, took: _Optional[float] = ..., results: _Optional[_Iterable[_Union[SearchResult, _Mapping]]] = ..., generative_grouped_result: _Optional[str] = ..., group_by_results: _Optional[_Iterable[_Union[GroupByResult, _Mapping]]] = ..., generative_grouped_results: _Optional[_Union[_generative_pb2.GenerativeResult, _Mapping]] = ...) -> None: ...
+    query_profile: QueryProfile
+    def __init__(self, took: _Optional[float] = ..., results: _Optional[_Iterable[_Union[SearchResult, _Mapping]]] = ..., generative_grouped_result: _Optional[str] = ..., group_by_results: _Optional[_Iterable[_Union[GroupByResult, _Mapping]]] = ..., generative_grouped_results: _Optional[_Union[_generative_pb2.GenerativeResult, _Mapping]] = ..., query_profile: _Optional[_Union[QueryProfile, _Mapping]] = ...) -> None: ...
+
+class QueryProfile(_message.Message):
+    __slots__ = ["shards"]
+    class SearchProfile(_message.Message):
+        __slots__ = ["details"]
+        class DetailsEntry(_message.Message):
+            __slots__ = ["key", "value"]
+            KEY_FIELD_NUMBER: _ClassVar[int]
+            VALUE_FIELD_NUMBER: _ClassVar[int]
+            key: str
+            value: str
+            def __init__(self, key: _Optional[str] = ..., value: _Optional[str] = ...) -> None: ...
+        DETAILS_FIELD_NUMBER: _ClassVar[int]
+        details: _containers.ScalarMap[str, str]
+        def __init__(self, details: _Optional[_Mapping[str, str]] = ...) -> None: ...
+    class ShardProfile(_message.Message):
+        __slots__ = ["name", "node", "searches"]
+        class SearchesEntry(_message.Message):
+            __slots__ = ["key", "value"]
+            KEY_FIELD_NUMBER: _ClassVar[int]
+            VALUE_FIELD_NUMBER: _ClassVar[int]
+            key: str
+            value: QueryProfile.SearchProfile
+            def __init__(self, key: _Optional[str] = ..., value: _Optional[_Union[QueryProfile.SearchProfile, _Mapping]] = ...) -> None: ...
+        NAME_FIELD_NUMBER: _ClassVar[int]
+        NODE_FIELD_NUMBER: _ClassVar[int]
+        SEARCHES_FIELD_NUMBER: _ClassVar[int]
+        name: str
+        node: str
+        searches: _containers.MessageMap[str, QueryProfile.SearchProfile]
+        def __init__(self, name: _Optional[str] = ..., node: _Optional[str] = ..., searches: _Optional[_Mapping[str, QueryProfile.SearchProfile]] = ...) -> None: ...
+    SHARDS_FIELD_NUMBER: _ClassVar[int]
+    shards: _containers.RepeatedCompositeFieldContainer[QueryProfile.ShardProfile]
+    def __init__(self, shards: _Optional[_Iterable[_Union[QueryProfile.ShardProfile, _Mapping]]] = ...) -> None: ...
 
 class RerankReply(_message.Message):
     __slots__ = ["score"]
diff --git a/weaviate/proto/v1/v5261/v1/search_get_pb2.py b/weaviate/proto/v1/v5261/v1/search_get_pb2.py
index 04356c19f..7686973af 100644
--- a/weaviate/proto/v1/v5261/v1/search_get_pb2.py
+++ b/weaviate/proto/v1/v5261/v1/search_get_pb2.py
@@ -18,7 +18,7 @@
 from weaviate.proto.v1.v5261.v1 import properties_pb2 as v1_dot_properties__pb2
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x13v1/search_get.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\x1a\x14v1/base_search.proto\x1a\x13v1/generative.proto\x1a\x13v1/properties.proto\"\x9c\x0b\n\rSearchRequest\x12\x12\n\ncollection\x18\x01 \x01(\t\x12\x0e\n\x06tenant\x18\n \x01(\t\x12=\n\x11\x63onsistency_level\x18\x0b \x01(\x0e\x32\x1d.weaviate.v1.ConsistencyLevelH\x00\x88\x01\x01\x12\x37\n\nproperties\x18\x14 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequestH\x01\x88\x01\x01\x12\x33\n\x08metadata\x18\x15 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequestH\x02\x88\x01\x01\x12+\n\x08group_by\x18\x16 \x01(\x0b\x32\x14.weaviate.v1.GroupByH\x03\x88\x01\x01\x12\r\n\x05limit\x18\x1e \x01(\r\x12\x0e\n\x06offset\x18\x1f \x01(\r\x12\x0f\n\x07\x61utocut\x18  \x01(\r\x12\r\n\x05\x61\x66ter\x18! \x01(\t\x12$\n\x07sort_by\x18\" \x03(\x0b\x32\x13.weaviate.v1.SortBy\x12*\n\x07\x66ilters\x18( \x01(\x0b\x32\x14.weaviate.v1.FiltersH\x04\x88\x01\x01\x12/\n\rhybrid_search\x18) \x01(\x0b\x32\x13.weaviate.v1.HybridH\x05\x88\x01\x01\x12+\n\x0b\x62m25_search\x18* \x01(\x0b\x32\x11.weaviate.v1.BM25H\x06\x88\x01\x01\x12\x31\n\x0bnear_vector\x18+ \x01(\x0b\x32\x17.weaviate.v1.NearVectorH\x07\x88\x01\x01\x12\x31\n\x0bnear_object\x18, \x01(\x0b\x32\x17.weaviate.v1.NearObjectH\x08\x88\x01\x01\x12\x33\n\tnear_text\x18- \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearchH\t\x88\x01\x01\x12\x35\n\nnear_image\x18. \x01(\x0b\x32\x1c.weaviate.v1.NearImageSearchH\n\x88\x01\x01\x12\x35\n\nnear_audio\x18/ \x01(\x0b\x32\x1c.weaviate.v1.NearAudioSearchH\x0b\x88\x01\x01\x12\x35\n\nnear_video\x18\x30 \x01(\x0b\x32\x1c.weaviate.v1.NearVideoSearchH\x0c\x88\x01\x01\x12\x35\n\nnear_depth\x18\x31 \x01(\x0b\x32\x1c.weaviate.v1.NearDepthSearchH\r\x88\x01\x01\x12\x39\n\x0cnear_thermal\x18\x32 \x01(\x0b\x32\x1e.weaviate.v1.NearThermalSearchH\x0e\x88\x01\x01\x12\x31\n\x08near_imu\x18\x33 \x01(\x0b\x32\x1a.weaviate.v1.NearIMUSearchH\x0f\x88\x01\x01\x12\x36\n\ngenerative\x18< \x01(\x0b\x32\x1d.weaviate.v1.GenerativeSearchH\x10\x88\x01\x01\x12(\n\x06rerank\x18= \x01(\x0b\x32\x13.weaviate.v1.RerankH\x11\x88\x01\x01\x12\x18\n\x0cuses_123_api\x18\x64 \x01(\x08\x42\x02\x18\x01\x12\x18\n\x0cuses_125_api\x18\x65 \x01(\x08\x42\x02\x18\x01\x12\x14\n\x0cuses_127_api\x18\x66 \x01(\x08\x42\x14\n\x12_consistency_levelB\r\n\x0b_propertiesB\x0b\n\t_metadataB\x0b\n\t_group_byB\n\n\x08_filtersB\x10\n\x0e_hybrid_searchB\x0e\n\x0c_bm25_searchB\x0e\n\x0c_near_vectorB\x0e\n\x0c_near_objectB\x0c\n\n_near_textB\r\n\x0b_near_imageB\r\n\x0b_near_audioB\r\n\x0b_near_videoB\r\n\x0b_near_depthB\x0f\n\r_near_thermalB\x0b\n\t_near_imuB\r\n\x0b_generativeB\t\n\x07_rerank\"L\n\x07GroupBy\x12\x0c\n\x04path\x18\x01 \x03(\t\x12\x18\n\x10number_of_groups\x18\x02 \x01(\x05\x12\x19\n\x11objects_per_group\x18\x03 \x01(\x05\")\n\x06SortBy\x12\x11\n\tascending\x18\x01 \x01(\x08\x12\x0c\n\x04path\x18\x02 \x03(\t\"\xdd\x01\n\x0fMetadataRequest\x12\x0c\n\x04uuid\x18\x01 \x01(\x08\x12\x0e\n\x06vector\x18\x02 \x01(\x08\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x04 \x01(\x08\x12\x10\n\x08\x64istance\x18\x05 \x01(\x08\x12\x11\n\tcertainty\x18\x06 \x01(\x08\x12\r\n\x05score\x18\x07 \x01(\x08\x12\x15\n\rexplain_score\x18\x08 \x01(\x08\x12\x15\n\ris_consistent\x18\t \x01(\x08\x12\x0f\n\x07vectors\x18\n \x03(\t\"\xd1\x01\n\x11PropertiesRequest\x12\x1a\n\x12non_ref_properties\x18\x01 \x03(\t\x12\x39\n\x0eref_properties\x18\x02 \x03(\x0b\x32!.weaviate.v1.RefPropertiesRequest\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\x12$\n\x1creturn_all_nonref_properties\x18\x0b \x01(\x08\"\x8b\x01\n\x17ObjectPropertiesRequest\x12\x11\n\tprop_name\x18\x01 \x01(\t\x12\x1c\n\x14primitive_properties\x18\x02 \x03(\t\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\"\xb1\x01\n\x14RefPropertiesRequest\x12\x1a\n\x12reference_property\x18\x01 \x01(\t\x12\x32\n\nproperties\x18\x02 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequest\x12.\n\x08metadata\x18\x03 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequest\x12\x19\n\x11target_collection\x18\x04 \x01(\t\"8\n\x06Rerank\x12\x10\n\x08property\x18\x01 \x01(\t\x12\x12\n\x05query\x18\x02 \x01(\tH\x00\x88\x01\x01\x42\x08\n\x06_query\"\xae\x02\n\x0bSearchReply\x12\x0c\n\x04took\x18\x01 \x01(\x02\x12*\n\x07results\x18\x02 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12*\n\x19generative_grouped_result\x18\x03 \x01(\tB\x02\x18\x01H\x00\x88\x01\x01\x12\x34\n\x10group_by_results\x18\x04 \x03(\x0b\x32\x1a.weaviate.v1.GroupByResult\x12\x46\n\x1agenerative_grouped_results\x18\x05 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x01\x88\x01\x01\x42\x1c\n\x1a_generative_grouped_resultB\x1d\n\x1b_generative_grouped_results\"\x1c\n\x0bRerankReply\x12\r\n\x05score\x18\x01 \x01(\x01\"\xe9\x02\n\rGroupByResult\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x14\n\x0cmin_distance\x18\x02 \x01(\x02\x12\x14\n\x0cmax_distance\x18\x03 \x01(\x02\x12\x19\n\x11number_of_objects\x18\x04 \x01(\x03\x12*\n\x07objects\x18\x05 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12-\n\x06rerank\x18\x06 \x01(\x0b\x32\x18.weaviate.v1.RerankReplyH\x00\x88\x01\x01\x12\x39\n\ngenerative\x18\x07 \x01(\x0b\x32\x1c.weaviate.v1.GenerativeReplyB\x02\x18\x01H\x01\x88\x01\x01\x12=\n\x11generative_result\x18\x08 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x02\x88\x01\x01\x42\t\n\x07_rerankB\r\n\x0b_generativeB\x14\n\x12_generative_result\"\xb7\x01\n\x0cSearchResult\x12\x31\n\nproperties\x18\x01 \x01(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12-\n\x08metadata\x18\x02 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12\x36\n\ngenerative\x18\x03 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x00\x88\x01\x01\x42\r\n\x0b_generative\"\xf7\x04\n\x0eMetadataResult\x12\n\n\x02id\x18\x01 \x01(\t\x12\x12\n\x06vector\x18\x02 \x03(\x02\x42\x02\x18\x01\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x03\x12\"\n\x1a\x63reation_time_unix_present\x18\x04 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x05 \x01(\x03\x12%\n\x1dlast_update_time_unix_present\x18\x06 \x01(\x08\x12\x10\n\x08\x64istance\x18\x07 \x01(\x02\x12\x18\n\x10\x64istance_present\x18\x08 \x01(\x08\x12\x11\n\tcertainty\x18\t \x01(\x02\x12\x19\n\x11\x63\x65rtainty_present\x18\n \x01(\x08\x12\r\n\x05score\x18\x0b \x01(\x02\x12\x15\n\rscore_present\x18\x0c \x01(\x08\x12\x15\n\rexplain_score\x18\r \x01(\t\x12\x1d\n\x15\x65xplain_score_present\x18\x0e \x01(\x08\x12\x1a\n\ris_consistent\x18\x0f \x01(\x08H\x00\x88\x01\x01\x12\x16\n\ngenerative\x18\x10 \x01(\tB\x02\x18\x01\x12\x1e\n\x12generative_present\x18\x11 \x01(\x08\x42\x02\x18\x01\x12\x1d\n\x15is_consistent_present\x18\x12 \x01(\x08\x12\x14\n\x0cvector_bytes\x18\x13 \x01(\x0c\x12\x13\n\x0bid_as_bytes\x18\x14 \x01(\x0c\x12\x14\n\x0crerank_score\x18\x15 \x01(\x01\x12\x1c\n\x14rerank_score_present\x18\x16 \x01(\x08\x12%\n\x07vectors\x18\x17 \x03(\x0b\x32\x14.weaviate.v1.VectorsB\x10\n\x0e_is_consistent\"\x88\x02\n\x10PropertiesResult\x12\x33\n\tref_props\x18\x02 \x03(\x0b\x32 .weaviate.v1.RefPropertiesResult\x12\x19\n\x11target_collection\x18\x03 \x01(\t\x12-\n\x08metadata\x18\x04 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12.\n\rnon_ref_props\x18\x0b \x01(\x0b\x32\x17.weaviate.v1.Properties\x12\x1b\n\x13ref_props_requested\x18\x0c \x01(\x08J\x04\x08\x01\x10\x02J\x04\x08\x05\x10\x06J\x04\x08\x06\x10\x07J\x04\x08\x07\x10\x08J\x04\x08\x08\x10\tJ\x04\x08\t\x10\nJ\x04\x08\n\x10\x0b\"[\n\x13RefPropertiesResult\x12\x31\n\nproperties\x18\x01 \x03(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12\x11\n\tprop_name\x18\x02 \x01(\tBs\n#io.weaviate.client.grpc.protocol.v1B\x16WeaviateProtoSearchGetZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x13v1/search_get.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\x1a\x14v1/base_search.proto\x1a\x13v1/generative.proto\x1a\x13v1/properties.proto\"\x9c\x0b\n\rSearchRequest\x12\x12\n\ncollection\x18\x01 \x01(\t\x12\x0e\n\x06tenant\x18\n \x01(\t\x12=\n\x11\x63onsistency_level\x18\x0b \x01(\x0e\x32\x1d.weaviate.v1.ConsistencyLevelH\x00\x88\x01\x01\x12\x37\n\nproperties\x18\x14 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequestH\x01\x88\x01\x01\x12\x33\n\x08metadata\x18\x15 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequestH\x02\x88\x01\x01\x12+\n\x08group_by\x18\x16 \x01(\x0b\x32\x14.weaviate.v1.GroupByH\x03\x88\x01\x01\x12\r\n\x05limit\x18\x1e \x01(\r\x12\x0e\n\x06offset\x18\x1f \x01(\r\x12\x0f\n\x07\x61utocut\x18  \x01(\r\x12\r\n\x05\x61\x66ter\x18! \x01(\t\x12$\n\x07sort_by\x18\" \x03(\x0b\x32\x13.weaviate.v1.SortBy\x12*\n\x07\x66ilters\x18( \x01(\x0b\x32\x14.weaviate.v1.FiltersH\x04\x88\x01\x01\x12/\n\rhybrid_search\x18) \x01(\x0b\x32\x13.weaviate.v1.HybridH\x05\x88\x01\x01\x12+\n\x0b\x62m25_search\x18* \x01(\x0b\x32\x11.weaviate.v1.BM25H\x06\x88\x01\x01\x12\x31\n\x0bnear_vector\x18+ \x01(\x0b\x32\x17.weaviate.v1.NearVectorH\x07\x88\x01\x01\x12\x31\n\x0bnear_object\x18, \x01(\x0b\x32\x17.weaviate.v1.NearObjectH\x08\x88\x01\x01\x12\x33\n\tnear_text\x18- \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearchH\t\x88\x01\x01\x12\x35\n\nnear_image\x18. \x01(\x0b\x32\x1c.weaviate.v1.NearImageSearchH\n\x88\x01\x01\x12\x35\n\nnear_audio\x18/ \x01(\x0b\x32\x1c.weaviate.v1.NearAudioSearchH\x0b\x88\x01\x01\x12\x35\n\nnear_video\x18\x30 \x01(\x0b\x32\x1c.weaviate.v1.NearVideoSearchH\x0c\x88\x01\x01\x12\x35\n\nnear_depth\x18\x31 \x01(\x0b\x32\x1c.weaviate.v1.NearDepthSearchH\r\x88\x01\x01\x12\x39\n\x0cnear_thermal\x18\x32 \x01(\x0b\x32\x1e.weaviate.v1.NearThermalSearchH\x0e\x88\x01\x01\x12\x31\n\x08near_imu\x18\x33 \x01(\x0b\x32\x1a.weaviate.v1.NearIMUSearchH\x0f\x88\x01\x01\x12\x36\n\ngenerative\x18< \x01(\x0b\x32\x1d.weaviate.v1.GenerativeSearchH\x10\x88\x01\x01\x12(\n\x06rerank\x18= \x01(\x0b\x32\x13.weaviate.v1.RerankH\x11\x88\x01\x01\x12\x18\n\x0cuses_123_api\x18\x64 \x01(\x08\x42\x02\x18\x01\x12\x18\n\x0cuses_125_api\x18\x65 \x01(\x08\x42\x02\x18\x01\x12\x14\n\x0cuses_127_api\x18\x66 \x01(\x08\x42\x14\n\x12_consistency_levelB\r\n\x0b_propertiesB\x0b\n\t_metadataB\x0b\n\t_group_byB\n\n\x08_filtersB\x10\n\x0e_hybrid_searchB\x0e\n\x0c_bm25_searchB\x0e\n\x0c_near_vectorB\x0e\n\x0c_near_objectB\x0c\n\n_near_textB\r\n\x0b_near_imageB\r\n\x0b_near_audioB\r\n\x0b_near_videoB\r\n\x0b_near_depthB\x0f\n\r_near_thermalB\x0b\n\t_near_imuB\r\n\x0b_generativeB\t\n\x07_rerank\"L\n\x07GroupBy\x12\x0c\n\x04path\x18\x01 \x03(\t\x12\x18\n\x10number_of_groups\x18\x02 \x01(\x05\x12\x19\n\x11objects_per_group\x18\x03 \x01(\x05\")\n\x06SortBy\x12\x11\n\tascending\x18\x01 \x01(\x08\x12\x0c\n\x04path\x18\x02 \x03(\t\"\xf4\x01\n\x0fMetadataRequest\x12\x0c\n\x04uuid\x18\x01 \x01(\x08\x12\x0e\n\x06vector\x18\x02 \x01(\x08\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x04 \x01(\x08\x12\x10\n\x08\x64istance\x18\x05 \x01(\x08\x12\x11\n\tcertainty\x18\x06 \x01(\x08\x12\r\n\x05score\x18\x07 \x01(\x08\x12\x15\n\rexplain_score\x18\x08 \x01(\x08\x12\x15\n\ris_consistent\x18\t \x01(\x08\x12\x0f\n\x07vectors\x18\n \x03(\t\x12\x15\n\rquery_profile\x18\x0b \x01(\x08\"\xd1\x01\n\x11PropertiesRequest\x12\x1a\n\x12non_ref_properties\x18\x01 \x03(\t\x12\x39\n\x0eref_properties\x18\x02 \x03(\x0b\x32!.weaviate.v1.RefPropertiesRequest\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\x12$\n\x1creturn_all_nonref_properties\x18\x0b \x01(\x08\"\x8b\x01\n\x17ObjectPropertiesRequest\x12\x11\n\tprop_name\x18\x01 \x01(\t\x12\x1c\n\x14primitive_properties\x18\x02 \x03(\t\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\"\xb1\x01\n\x14RefPropertiesRequest\x12\x1a\n\x12reference_property\x18\x01 \x01(\t\x12\x32\n\nproperties\x18\x02 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequest\x12.\n\x08metadata\x18\x03 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequest\x12\x19\n\x11target_collection\x18\x04 \x01(\t\"8\n\x06Rerank\x12\x10\n\x08property\x18\x01 \x01(\t\x12\x12\n\x05query\x18\x02 \x01(\tH\x00\x88\x01\x01\x42\x08\n\x06_query\"\xf7\x02\n\x0bSearchReply\x12\x0c\n\x04took\x18\x01 \x01(\x02\x12*\n\x07results\x18\x02 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12*\n\x19generative_grouped_result\x18\x03 \x01(\tB\x02\x18\x01H\x00\x88\x01\x01\x12\x34\n\x10group_by_results\x18\x04 \x03(\x0b\x32\x1a.weaviate.v1.GroupByResult\x12\x46\n\x1agenerative_grouped_results\x18\x05 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x01\x88\x01\x01\x12\x35\n\rquery_profile\x18\x06 \x01(\x0b\x32\x19.weaviate.v1.QueryProfileH\x02\x88\x01\x01\x42\x1c\n\x1a_generative_grouped_resultB\x1d\n\x1b_generative_grouped_resultsB\x10\n\x0e_query_profile\"\x9e\x03\n\x0cQueryProfile\x12\x36\n\x06shards\x18\x01 \x03(\x0b\x32&.weaviate.v1.QueryProfile.ShardProfile\x1a\x86\x01\n\rSearchProfile\x12\x45\n\x07\x64\x65tails\x18\x01 \x03(\x0b\x32\x34.weaviate.v1.QueryProfile.SearchProfile.DetailsEntry\x1a.\n\x0c\x44\x65tailsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01\x1a\xcc\x01\n\x0cShardProfile\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x0c\n\x04node\x18\x02 \x01(\t\x12\x46\n\x08searches\x18\x03 \x03(\x0b\x32\x34.weaviate.v1.QueryProfile.ShardProfile.SearchesEntry\x1aX\n\rSearchesEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\x36\n\x05value\x18\x02 \x01(\x0b\x32\'.weaviate.v1.QueryProfile.SearchProfile:\x02\x38\x01\"\x1c\n\x0bRerankReply\x12\r\n\x05score\x18\x01 \x01(\x01\"\xe9\x02\n\rGroupByResult\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x14\n\x0cmin_distance\x18\x02 \x01(\x02\x12\x14\n\x0cmax_distance\x18\x03 \x01(\x02\x12\x19\n\x11number_of_objects\x18\x04 \x01(\x03\x12*\n\x07objects\x18\x05 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12-\n\x06rerank\x18\x06 \x01(\x0b\x32\x18.weaviate.v1.RerankReplyH\x00\x88\x01\x01\x12\x39\n\ngenerative\x18\x07 \x01(\x0b\x32\x1c.weaviate.v1.GenerativeReplyB\x02\x18\x01H\x01\x88\x01\x01\x12=\n\x11generative_result\x18\x08 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x02\x88\x01\x01\x42\t\n\x07_rerankB\r\n\x0b_generativeB\x14\n\x12_generative_result\"\xb7\x01\n\x0cSearchResult\x12\x31\n\nproperties\x18\x01 \x01(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12-\n\x08metadata\x18\x02 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12\x36\n\ngenerative\x18\x03 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x00\x88\x01\x01\x42\r\n\x0b_generative\"\xf7\x04\n\x0eMetadataResult\x12\n\n\x02id\x18\x01 \x01(\t\x12\x12\n\x06vector\x18\x02 \x03(\x02\x42\x02\x18\x01\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x03\x12\"\n\x1a\x63reation_time_unix_present\x18\x04 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x05 \x01(\x03\x12%\n\x1dlast_update_time_unix_present\x18\x06 \x01(\x08\x12\x10\n\x08\x64istance\x18\x07 \x01(\x02\x12\x18\n\x10\x64istance_present\x18\x08 \x01(\x08\x12\x11\n\tcertainty\x18\t \x01(\x02\x12\x19\n\x11\x63\x65rtainty_present\x18\n \x01(\x08\x12\r\n\x05score\x18\x0b \x01(\x02\x12\x15\n\rscore_present\x18\x0c \x01(\x08\x12\x15\n\rexplain_score\x18\r \x01(\t\x12\x1d\n\x15\x65xplain_score_present\x18\x0e \x01(\x08\x12\x1a\n\ris_consistent\x18\x0f \x01(\x08H\x00\x88\x01\x01\x12\x16\n\ngenerative\x18\x10 \x01(\tB\x02\x18\x01\x12\x1e\n\x12generative_present\x18\x11 \x01(\x08\x42\x02\x18\x01\x12\x1d\n\x15is_consistent_present\x18\x12 \x01(\x08\x12\x14\n\x0cvector_bytes\x18\x13 \x01(\x0c\x12\x13\n\x0bid_as_bytes\x18\x14 \x01(\x0c\x12\x14\n\x0crerank_score\x18\x15 \x01(\x01\x12\x1c\n\x14rerank_score_present\x18\x16 \x01(\x08\x12%\n\x07vectors\x18\x17 \x03(\x0b\x32\x14.weaviate.v1.VectorsB\x10\n\x0e_is_consistent\"\x88\x02\n\x10PropertiesResult\x12\x33\n\tref_props\x18\x02 \x03(\x0b\x32 .weaviate.v1.RefPropertiesResult\x12\x19\n\x11target_collection\x18\x03 \x01(\t\x12-\n\x08metadata\x18\x04 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12.\n\rnon_ref_props\x18\x0b \x01(\x0b\x32\x17.weaviate.v1.Properties\x12\x1b\n\x13ref_props_requested\x18\x0c \x01(\x08J\x04\x08\x01\x10\x02J\x04\x08\x05\x10\x06J\x04\x08\x06\x10\x07J\x04\x08\x07\x10\x08J\x04\x08\x08\x10\tJ\x04\x08\t\x10\nJ\x04\x08\n\x10\x0b\"[\n\x13RefPropertiesResult\x12\x31\n\nproperties\x18\x01 \x03(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12\x11\n\tprop_name\x18\x02 \x01(\tBs\n#io.weaviate.client.grpc.protocol.v1B\x16WeaviateProtoSearchGetZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -32,6 +32,10 @@
   _globals['_SEARCHREQUEST'].fields_by_name['uses_125_api']._serialized_options = b'\030\001'
   _globals['_SEARCHREPLY'].fields_by_name['generative_grouped_result']._loaded_options = None
   _globals['_SEARCHREPLY'].fields_by_name['generative_grouped_result']._serialized_options = b'\030\001'
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._loaded_options = None
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._serialized_options = b'8\001'
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._loaded_options = None
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._serialized_options = b'8\001'
   _globals['_GROUPBYRESULT'].fields_by_name['generative']._loaded_options = None
   _globals['_GROUPBYRESULT'].fields_by_name['generative']._serialized_options = b'\030\001'
   _globals['_METADATARESULT'].fields_by_name['vector']._loaded_options = None
@@ -47,27 +51,37 @@
   _globals['_SORTBY']._serialized_start=1632
   _globals['_SORTBY']._serialized_end=1673
   _globals['_METADATAREQUEST']._serialized_start=1676
-  _globals['_METADATAREQUEST']._serialized_end=1897
-  _globals['_PROPERTIESREQUEST']._serialized_start=1900
-  _globals['_PROPERTIESREQUEST']._serialized_end=2109
-  _globals['_OBJECTPROPERTIESREQUEST']._serialized_start=2112
-  _globals['_OBJECTPROPERTIESREQUEST']._serialized_end=2251
-  _globals['_REFPROPERTIESREQUEST']._serialized_start=2254
-  _globals['_REFPROPERTIESREQUEST']._serialized_end=2431
-  _globals['_RERANK']._serialized_start=2433
-  _globals['_RERANK']._serialized_end=2489
-  _globals['_SEARCHREPLY']._serialized_start=2492
-  _globals['_SEARCHREPLY']._serialized_end=2794
-  _globals['_RERANKREPLY']._serialized_start=2796
-  _globals['_RERANKREPLY']._serialized_end=2824
-  _globals['_GROUPBYRESULT']._serialized_start=2827
-  _globals['_GROUPBYRESULT']._serialized_end=3188
-  _globals['_SEARCHRESULT']._serialized_start=3191
-  _globals['_SEARCHRESULT']._serialized_end=3374
-  _globals['_METADATARESULT']._serialized_start=3377
-  _globals['_METADATARESULT']._serialized_end=4008
-  _globals['_PROPERTIESRESULT']._serialized_start=4011
-  _globals['_PROPERTIESRESULT']._serialized_end=4275
-  _globals['_REFPROPERTIESRESULT']._serialized_start=4277
-  _globals['_REFPROPERTIESRESULT']._serialized_end=4368
+  _globals['_METADATAREQUEST']._serialized_end=1920
+  _globals['_PROPERTIESREQUEST']._serialized_start=1923
+  _globals['_PROPERTIESREQUEST']._serialized_end=2132
+  _globals['_OBJECTPROPERTIESREQUEST']._serialized_start=2135
+  _globals['_OBJECTPROPERTIESREQUEST']._serialized_end=2274
+  _globals['_REFPROPERTIESREQUEST']._serialized_start=2277
+  _globals['_REFPROPERTIESREQUEST']._serialized_end=2454
+  _globals['_RERANK']._serialized_start=2456
+  _globals['_RERANK']._serialized_end=2512
+  _globals['_SEARCHREPLY']._serialized_start=2515
+  _globals['_SEARCHREPLY']._serialized_end=2890
+  _globals['_QUERYPROFILE']._serialized_start=2893
+  _globals['_QUERYPROFILE']._serialized_end=3307
+  _globals['_QUERYPROFILE_SEARCHPROFILE']._serialized_start=2966
+  _globals['_QUERYPROFILE_SEARCHPROFILE']._serialized_end=3100
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._serialized_start=3054
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._serialized_end=3100
+  _globals['_QUERYPROFILE_SHARDPROFILE']._serialized_start=3103
+  _globals['_QUERYPROFILE_SHARDPROFILE']._serialized_end=3307
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._serialized_start=3219
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._serialized_end=3307
+  _globals['_RERANKREPLY']._serialized_start=3309
+  _globals['_RERANKREPLY']._serialized_end=3337
+  _globals['_GROUPBYRESULT']._serialized_start=3340
+  _globals['_GROUPBYRESULT']._serialized_end=3701
+  _globals['_SEARCHRESULT']._serialized_start=3704
+  _globals['_SEARCHRESULT']._serialized_end=3887
+  _globals['_METADATARESULT']._serialized_start=3890
+  _globals['_METADATARESULT']._serialized_end=4521
+  _globals['_PROPERTIESRESULT']._serialized_start=4524
+  _globals['_PROPERTIESRESULT']._serialized_end=4788
+  _globals['_REFPROPERTIESRESULT']._serialized_start=4790
+  _globals['_REFPROPERTIESRESULT']._serialized_end=4881
 # @@protoc_insertion_point(module_scope)
diff --git a/weaviate/proto/v1/v5261/v1/search_get_pb2.pyi b/weaviate/proto/v1/v5261/v1/search_get_pb2.pyi
index 4a28237d9..783ca3564 100644
--- a/weaviate/proto/v1/v5261/v1/search_get_pb2.pyi
+++ b/weaviate/proto/v1/v5261/v1/search_get_pb2.pyi
@@ -88,7 +88,7 @@ class SortBy(_message.Message):
     def __init__(self, ascending: bool = ..., path: _Optional[_Iterable[str]] = ...) -> None: ...
 
 class MetadataRequest(_message.Message):
-    __slots__ = ("uuid", "vector", "creation_time_unix", "last_update_time_unix", "distance", "certainty", "score", "explain_score", "is_consistent", "vectors")
+    __slots__ = ("uuid", "vector", "creation_time_unix", "last_update_time_unix", "distance", "certainty", "score", "explain_score", "is_consistent", "vectors", "query_profile")
     UUID_FIELD_NUMBER: _ClassVar[int]
     VECTOR_FIELD_NUMBER: _ClassVar[int]
     CREATION_TIME_UNIX_FIELD_NUMBER: _ClassVar[int]
@@ -99,6 +99,7 @@ class MetadataRequest(_message.Message):
     EXPLAIN_SCORE_FIELD_NUMBER: _ClassVar[int]
     IS_CONSISTENT_FIELD_NUMBER: _ClassVar[int]
     VECTORS_FIELD_NUMBER: _ClassVar[int]
+    QUERY_PROFILE_FIELD_NUMBER: _ClassVar[int]
     uuid: bool
     vector: bool
     creation_time_unix: bool
@@ -109,7 +110,8 @@ class MetadataRequest(_message.Message):
     explain_score: bool
     is_consistent: bool
     vectors: _containers.RepeatedScalarFieldContainer[str]
-    def __init__(self, uuid: bool = ..., vector: bool = ..., creation_time_unix: bool = ..., last_update_time_unix: bool = ..., distance: bool = ..., certainty: bool = ..., score: bool = ..., explain_score: bool = ..., is_consistent: bool = ..., vectors: _Optional[_Iterable[str]] = ...) -> None: ...
+    query_profile: bool
+    def __init__(self, uuid: bool = ..., vector: bool = ..., creation_time_unix: bool = ..., last_update_time_unix: bool = ..., distance: bool = ..., certainty: bool = ..., score: bool = ..., explain_score: bool = ..., is_consistent: bool = ..., vectors: _Optional[_Iterable[str]] = ..., query_profile: bool = ...) -> None: ...
 
 class PropertiesRequest(_message.Message):
     __slots__ = ("non_ref_properties", "ref_properties", "object_properties", "return_all_nonref_properties")
@@ -154,18 +156,54 @@ class Rerank(_message.Message):
     def __init__(self, property: _Optional[str] = ..., query: _Optional[str] = ...) -> None: ...
 
 class SearchReply(_message.Message):
-    __slots__ = ("took", "results", "generative_grouped_result", "group_by_results", "generative_grouped_results")
+    __slots__ = ("took", "results", "generative_grouped_result", "group_by_results", "generative_grouped_results", "query_profile")
     TOOK_FIELD_NUMBER: _ClassVar[int]
     RESULTS_FIELD_NUMBER: _ClassVar[int]
     GENERATIVE_GROUPED_RESULT_FIELD_NUMBER: _ClassVar[int]
     GROUP_BY_RESULTS_FIELD_NUMBER: _ClassVar[int]
     GENERATIVE_GROUPED_RESULTS_FIELD_NUMBER: _ClassVar[int]
+    QUERY_PROFILE_FIELD_NUMBER: _ClassVar[int]
     took: float
     results: _containers.RepeatedCompositeFieldContainer[SearchResult]
     generative_grouped_result: str
     group_by_results: _containers.RepeatedCompositeFieldContainer[GroupByResult]
     generative_grouped_results: _generative_pb2.GenerativeResult
-    def __init__(self, took: _Optional[float] = ..., results: _Optional[_Iterable[_Union[SearchResult, _Mapping]]] = ..., generative_grouped_result: _Optional[str] = ..., group_by_results: _Optional[_Iterable[_Union[GroupByResult, _Mapping]]] = ..., generative_grouped_results: _Optional[_Union[_generative_pb2.GenerativeResult, _Mapping]] = ...) -> None: ...
+    query_profile: QueryProfile
+    def __init__(self, took: _Optional[float] = ..., results: _Optional[_Iterable[_Union[SearchResult, _Mapping]]] = ..., generative_grouped_result: _Optional[str] = ..., group_by_results: _Optional[_Iterable[_Union[GroupByResult, _Mapping]]] = ..., generative_grouped_results: _Optional[_Union[_generative_pb2.GenerativeResult, _Mapping]] = ..., query_profile: _Optional[_Union[QueryProfile, _Mapping]] = ...) -> None: ...
+
+class QueryProfile(_message.Message):
+    __slots__ = ("shards",)
+    class SearchProfile(_message.Message):
+        __slots__ = ("details",)
+        class DetailsEntry(_message.Message):
+            __slots__ = ("key", "value")
+            KEY_FIELD_NUMBER: _ClassVar[int]
+            VALUE_FIELD_NUMBER: _ClassVar[int]
+            key: str
+            value: str
+            def __init__(self, key: _Optional[str] = ..., value: _Optional[str] = ...) -> None: ...
+        DETAILS_FIELD_NUMBER: _ClassVar[int]
+        details: _containers.ScalarMap[str, str]
+        def __init__(self, details: _Optional[_Mapping[str, str]] = ...) -> None: ...
+    class ShardProfile(_message.Message):
+        __slots__ = ("name", "node", "searches")
+        class SearchesEntry(_message.Message):
+            __slots__ = ("key", "value")
+            KEY_FIELD_NUMBER: _ClassVar[int]
+            VALUE_FIELD_NUMBER: _ClassVar[int]
+            key: str
+            value: QueryProfile.SearchProfile
+            def __init__(self, key: _Optional[str] = ..., value: _Optional[_Union[QueryProfile.SearchProfile, _Mapping]] = ...) -> None: ...
+        NAME_FIELD_NUMBER: _ClassVar[int]
+        NODE_FIELD_NUMBER: _ClassVar[int]
+        SEARCHES_FIELD_NUMBER: _ClassVar[int]
+        name: str
+        node: str
+        searches: _containers.MessageMap[str, QueryProfile.SearchProfile]
+        def __init__(self, name: _Optional[str] = ..., node: _Optional[str] = ..., searches: _Optional[_Mapping[str, QueryProfile.SearchProfile]] = ...) -> None: ...
+    SHARDS_FIELD_NUMBER: _ClassVar[int]
+    shards: _containers.RepeatedCompositeFieldContainer[QueryProfile.ShardProfile]
+    def __init__(self, shards: _Optional[_Iterable[_Union[QueryProfile.ShardProfile, _Mapping]]] = ...) -> None: ...
 
 class RerankReply(_message.Message):
     __slots__ = ("score",)
diff --git a/weaviate/proto/v1/v6300/v1/search_get_pb2.py b/weaviate/proto/v1/v6300/v1/search_get_pb2.py
index 88dfb5992..1b38e4fb6 100644
--- a/weaviate/proto/v1/v6300/v1/search_get_pb2.py
+++ b/weaviate/proto/v1/v6300/v1/search_get_pb2.py
@@ -28,7 +28,7 @@
 from weaviate.proto.v1.v6300.v1 import properties_pb2 as v1_dot_properties__pb2
 
 
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x13v1/search_get.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\x1a\x14v1/base_search.proto\x1a\x13v1/generative.proto\x1a\x13v1/properties.proto\"\x9c\x0b\n\rSearchRequest\x12\x12\n\ncollection\x18\x01 \x01(\t\x12\x0e\n\x06tenant\x18\n \x01(\t\x12=\n\x11\x63onsistency_level\x18\x0b \x01(\x0e\x32\x1d.weaviate.v1.ConsistencyLevelH\x00\x88\x01\x01\x12\x37\n\nproperties\x18\x14 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequestH\x01\x88\x01\x01\x12\x33\n\x08metadata\x18\x15 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequestH\x02\x88\x01\x01\x12+\n\x08group_by\x18\x16 \x01(\x0b\x32\x14.weaviate.v1.GroupByH\x03\x88\x01\x01\x12\r\n\x05limit\x18\x1e \x01(\r\x12\x0e\n\x06offset\x18\x1f \x01(\r\x12\x0f\n\x07\x61utocut\x18  \x01(\r\x12\r\n\x05\x61\x66ter\x18! \x01(\t\x12$\n\x07sort_by\x18\" \x03(\x0b\x32\x13.weaviate.v1.SortBy\x12*\n\x07\x66ilters\x18( \x01(\x0b\x32\x14.weaviate.v1.FiltersH\x04\x88\x01\x01\x12/\n\rhybrid_search\x18) \x01(\x0b\x32\x13.weaviate.v1.HybridH\x05\x88\x01\x01\x12+\n\x0b\x62m25_search\x18* \x01(\x0b\x32\x11.weaviate.v1.BM25H\x06\x88\x01\x01\x12\x31\n\x0bnear_vector\x18+ \x01(\x0b\x32\x17.weaviate.v1.NearVectorH\x07\x88\x01\x01\x12\x31\n\x0bnear_object\x18, \x01(\x0b\x32\x17.weaviate.v1.NearObjectH\x08\x88\x01\x01\x12\x33\n\tnear_text\x18- \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearchH\t\x88\x01\x01\x12\x35\n\nnear_image\x18. \x01(\x0b\x32\x1c.weaviate.v1.NearImageSearchH\n\x88\x01\x01\x12\x35\n\nnear_audio\x18/ \x01(\x0b\x32\x1c.weaviate.v1.NearAudioSearchH\x0b\x88\x01\x01\x12\x35\n\nnear_video\x18\x30 \x01(\x0b\x32\x1c.weaviate.v1.NearVideoSearchH\x0c\x88\x01\x01\x12\x35\n\nnear_depth\x18\x31 \x01(\x0b\x32\x1c.weaviate.v1.NearDepthSearchH\r\x88\x01\x01\x12\x39\n\x0cnear_thermal\x18\x32 \x01(\x0b\x32\x1e.weaviate.v1.NearThermalSearchH\x0e\x88\x01\x01\x12\x31\n\x08near_imu\x18\x33 \x01(\x0b\x32\x1a.weaviate.v1.NearIMUSearchH\x0f\x88\x01\x01\x12\x36\n\ngenerative\x18< \x01(\x0b\x32\x1d.weaviate.v1.GenerativeSearchH\x10\x88\x01\x01\x12(\n\x06rerank\x18= \x01(\x0b\x32\x13.weaviate.v1.RerankH\x11\x88\x01\x01\x12\x18\n\x0cuses_123_api\x18\x64 \x01(\x08\x42\x02\x18\x01\x12\x18\n\x0cuses_125_api\x18\x65 \x01(\x08\x42\x02\x18\x01\x12\x14\n\x0cuses_127_api\x18\x66 \x01(\x08\x42\x14\n\x12_consistency_levelB\r\n\x0b_propertiesB\x0b\n\t_metadataB\x0b\n\t_group_byB\n\n\x08_filtersB\x10\n\x0e_hybrid_searchB\x0e\n\x0c_bm25_searchB\x0e\n\x0c_near_vectorB\x0e\n\x0c_near_objectB\x0c\n\n_near_textB\r\n\x0b_near_imageB\r\n\x0b_near_audioB\r\n\x0b_near_videoB\r\n\x0b_near_depthB\x0f\n\r_near_thermalB\x0b\n\t_near_imuB\r\n\x0b_generativeB\t\n\x07_rerank\"L\n\x07GroupBy\x12\x0c\n\x04path\x18\x01 \x03(\t\x12\x18\n\x10number_of_groups\x18\x02 \x01(\x05\x12\x19\n\x11objects_per_group\x18\x03 \x01(\x05\")\n\x06SortBy\x12\x11\n\tascending\x18\x01 \x01(\x08\x12\x0c\n\x04path\x18\x02 \x03(\t\"\xdd\x01\n\x0fMetadataRequest\x12\x0c\n\x04uuid\x18\x01 \x01(\x08\x12\x0e\n\x06vector\x18\x02 \x01(\x08\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x04 \x01(\x08\x12\x10\n\x08\x64istance\x18\x05 \x01(\x08\x12\x11\n\tcertainty\x18\x06 \x01(\x08\x12\r\n\x05score\x18\x07 \x01(\x08\x12\x15\n\rexplain_score\x18\x08 \x01(\x08\x12\x15\n\ris_consistent\x18\t \x01(\x08\x12\x0f\n\x07vectors\x18\n \x03(\t\"\xd1\x01\n\x11PropertiesRequest\x12\x1a\n\x12non_ref_properties\x18\x01 \x03(\t\x12\x39\n\x0eref_properties\x18\x02 \x03(\x0b\x32!.weaviate.v1.RefPropertiesRequest\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\x12$\n\x1creturn_all_nonref_properties\x18\x0b \x01(\x08\"\x8b\x01\n\x17ObjectPropertiesRequest\x12\x11\n\tprop_name\x18\x01 \x01(\t\x12\x1c\n\x14primitive_properties\x18\x02 \x03(\t\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\"\xb1\x01\n\x14RefPropertiesRequest\x12\x1a\n\x12reference_property\x18\x01 \x01(\t\x12\x32\n\nproperties\x18\x02 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequest\x12.\n\x08metadata\x18\x03 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequest\x12\x19\n\x11target_collection\x18\x04 \x01(\t\"8\n\x06Rerank\x12\x10\n\x08property\x18\x01 \x01(\t\x12\x12\n\x05query\x18\x02 \x01(\tH\x00\x88\x01\x01\x42\x08\n\x06_query\"\xae\x02\n\x0bSearchReply\x12\x0c\n\x04took\x18\x01 \x01(\x02\x12*\n\x07results\x18\x02 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12*\n\x19generative_grouped_result\x18\x03 \x01(\tB\x02\x18\x01H\x00\x88\x01\x01\x12\x34\n\x10group_by_results\x18\x04 \x03(\x0b\x32\x1a.weaviate.v1.GroupByResult\x12\x46\n\x1agenerative_grouped_results\x18\x05 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x01\x88\x01\x01\x42\x1c\n\x1a_generative_grouped_resultB\x1d\n\x1b_generative_grouped_results\"\x1c\n\x0bRerankReply\x12\r\n\x05score\x18\x01 \x01(\x01\"\xe9\x02\n\rGroupByResult\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x14\n\x0cmin_distance\x18\x02 \x01(\x02\x12\x14\n\x0cmax_distance\x18\x03 \x01(\x02\x12\x19\n\x11number_of_objects\x18\x04 \x01(\x03\x12*\n\x07objects\x18\x05 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12-\n\x06rerank\x18\x06 \x01(\x0b\x32\x18.weaviate.v1.RerankReplyH\x00\x88\x01\x01\x12\x39\n\ngenerative\x18\x07 \x01(\x0b\x32\x1c.weaviate.v1.GenerativeReplyB\x02\x18\x01H\x01\x88\x01\x01\x12=\n\x11generative_result\x18\x08 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x02\x88\x01\x01\x42\t\n\x07_rerankB\r\n\x0b_generativeB\x14\n\x12_generative_result\"\xb7\x01\n\x0cSearchResult\x12\x31\n\nproperties\x18\x01 \x01(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12-\n\x08metadata\x18\x02 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12\x36\n\ngenerative\x18\x03 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x00\x88\x01\x01\x42\r\n\x0b_generative\"\xf7\x04\n\x0eMetadataResult\x12\n\n\x02id\x18\x01 \x01(\t\x12\x12\n\x06vector\x18\x02 \x03(\x02\x42\x02\x18\x01\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x03\x12\"\n\x1a\x63reation_time_unix_present\x18\x04 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x05 \x01(\x03\x12%\n\x1dlast_update_time_unix_present\x18\x06 \x01(\x08\x12\x10\n\x08\x64istance\x18\x07 \x01(\x02\x12\x18\n\x10\x64istance_present\x18\x08 \x01(\x08\x12\x11\n\tcertainty\x18\t \x01(\x02\x12\x19\n\x11\x63\x65rtainty_present\x18\n \x01(\x08\x12\r\n\x05score\x18\x0b \x01(\x02\x12\x15\n\rscore_present\x18\x0c \x01(\x08\x12\x15\n\rexplain_score\x18\r \x01(\t\x12\x1d\n\x15\x65xplain_score_present\x18\x0e \x01(\x08\x12\x1a\n\ris_consistent\x18\x0f \x01(\x08H\x00\x88\x01\x01\x12\x16\n\ngenerative\x18\x10 \x01(\tB\x02\x18\x01\x12\x1e\n\x12generative_present\x18\x11 \x01(\x08\x42\x02\x18\x01\x12\x1d\n\x15is_consistent_present\x18\x12 \x01(\x08\x12\x14\n\x0cvector_bytes\x18\x13 \x01(\x0c\x12\x13\n\x0bid_as_bytes\x18\x14 \x01(\x0c\x12\x14\n\x0crerank_score\x18\x15 \x01(\x01\x12\x1c\n\x14rerank_score_present\x18\x16 \x01(\x08\x12%\n\x07vectors\x18\x17 \x03(\x0b\x32\x14.weaviate.v1.VectorsB\x10\n\x0e_is_consistent\"\x88\x02\n\x10PropertiesResult\x12\x33\n\tref_props\x18\x02 \x03(\x0b\x32 .weaviate.v1.RefPropertiesResult\x12\x19\n\x11target_collection\x18\x03 \x01(\t\x12-\n\x08metadata\x18\x04 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12.\n\rnon_ref_props\x18\x0b \x01(\x0b\x32\x17.weaviate.v1.Properties\x12\x1b\n\x13ref_props_requested\x18\x0c \x01(\x08J\x04\x08\x01\x10\x02J\x04\x08\x05\x10\x06J\x04\x08\x06\x10\x07J\x04\x08\x07\x10\x08J\x04\x08\x08\x10\tJ\x04\x08\t\x10\nJ\x04\x08\n\x10\x0b\"[\n\x13RefPropertiesResult\x12\x31\n\nproperties\x18\x01 \x03(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12\x11\n\tprop_name\x18\x02 \x01(\tBs\n#io.weaviate.client.grpc.protocol.v1B\x16WeaviateProtoSearchGetZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\x13v1/search_get.proto\x12\x0bweaviate.v1\x1a\rv1/base.proto\x1a\x14v1/base_search.proto\x1a\x13v1/generative.proto\x1a\x13v1/properties.proto\"\x9c\x0b\n\rSearchRequest\x12\x12\n\ncollection\x18\x01 \x01(\t\x12\x0e\n\x06tenant\x18\n \x01(\t\x12=\n\x11\x63onsistency_level\x18\x0b \x01(\x0e\x32\x1d.weaviate.v1.ConsistencyLevelH\x00\x88\x01\x01\x12\x37\n\nproperties\x18\x14 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequestH\x01\x88\x01\x01\x12\x33\n\x08metadata\x18\x15 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequestH\x02\x88\x01\x01\x12+\n\x08group_by\x18\x16 \x01(\x0b\x32\x14.weaviate.v1.GroupByH\x03\x88\x01\x01\x12\r\n\x05limit\x18\x1e \x01(\r\x12\x0e\n\x06offset\x18\x1f \x01(\r\x12\x0f\n\x07\x61utocut\x18  \x01(\r\x12\r\n\x05\x61\x66ter\x18! \x01(\t\x12$\n\x07sort_by\x18\" \x03(\x0b\x32\x13.weaviate.v1.SortBy\x12*\n\x07\x66ilters\x18( \x01(\x0b\x32\x14.weaviate.v1.FiltersH\x04\x88\x01\x01\x12/\n\rhybrid_search\x18) \x01(\x0b\x32\x13.weaviate.v1.HybridH\x05\x88\x01\x01\x12+\n\x0b\x62m25_search\x18* \x01(\x0b\x32\x11.weaviate.v1.BM25H\x06\x88\x01\x01\x12\x31\n\x0bnear_vector\x18+ \x01(\x0b\x32\x17.weaviate.v1.NearVectorH\x07\x88\x01\x01\x12\x31\n\x0bnear_object\x18, \x01(\x0b\x32\x17.weaviate.v1.NearObjectH\x08\x88\x01\x01\x12\x33\n\tnear_text\x18- \x01(\x0b\x32\x1b.weaviate.v1.NearTextSearchH\t\x88\x01\x01\x12\x35\n\nnear_image\x18. \x01(\x0b\x32\x1c.weaviate.v1.NearImageSearchH\n\x88\x01\x01\x12\x35\n\nnear_audio\x18/ \x01(\x0b\x32\x1c.weaviate.v1.NearAudioSearchH\x0b\x88\x01\x01\x12\x35\n\nnear_video\x18\x30 \x01(\x0b\x32\x1c.weaviate.v1.NearVideoSearchH\x0c\x88\x01\x01\x12\x35\n\nnear_depth\x18\x31 \x01(\x0b\x32\x1c.weaviate.v1.NearDepthSearchH\r\x88\x01\x01\x12\x39\n\x0cnear_thermal\x18\x32 \x01(\x0b\x32\x1e.weaviate.v1.NearThermalSearchH\x0e\x88\x01\x01\x12\x31\n\x08near_imu\x18\x33 \x01(\x0b\x32\x1a.weaviate.v1.NearIMUSearchH\x0f\x88\x01\x01\x12\x36\n\ngenerative\x18< \x01(\x0b\x32\x1d.weaviate.v1.GenerativeSearchH\x10\x88\x01\x01\x12(\n\x06rerank\x18= \x01(\x0b\x32\x13.weaviate.v1.RerankH\x11\x88\x01\x01\x12\x18\n\x0cuses_123_api\x18\x64 \x01(\x08\x42\x02\x18\x01\x12\x18\n\x0cuses_125_api\x18\x65 \x01(\x08\x42\x02\x18\x01\x12\x14\n\x0cuses_127_api\x18\x66 \x01(\x08\x42\x14\n\x12_consistency_levelB\r\n\x0b_propertiesB\x0b\n\t_metadataB\x0b\n\t_group_byB\n\n\x08_filtersB\x10\n\x0e_hybrid_searchB\x0e\n\x0c_bm25_searchB\x0e\n\x0c_near_vectorB\x0e\n\x0c_near_objectB\x0c\n\n_near_textB\r\n\x0b_near_imageB\r\n\x0b_near_audioB\r\n\x0b_near_videoB\r\n\x0b_near_depthB\x0f\n\r_near_thermalB\x0b\n\t_near_imuB\r\n\x0b_generativeB\t\n\x07_rerank\"L\n\x07GroupBy\x12\x0c\n\x04path\x18\x01 \x03(\t\x12\x18\n\x10number_of_groups\x18\x02 \x01(\x05\x12\x19\n\x11objects_per_group\x18\x03 \x01(\x05\")\n\x06SortBy\x12\x11\n\tascending\x18\x01 \x01(\x08\x12\x0c\n\x04path\x18\x02 \x03(\t\"\xf4\x01\n\x0fMetadataRequest\x12\x0c\n\x04uuid\x18\x01 \x01(\x08\x12\x0e\n\x06vector\x18\x02 \x01(\x08\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x04 \x01(\x08\x12\x10\n\x08\x64istance\x18\x05 \x01(\x08\x12\x11\n\tcertainty\x18\x06 \x01(\x08\x12\r\n\x05score\x18\x07 \x01(\x08\x12\x15\n\rexplain_score\x18\x08 \x01(\x08\x12\x15\n\ris_consistent\x18\t \x01(\x08\x12\x0f\n\x07vectors\x18\n \x03(\t\x12\x15\n\rquery_profile\x18\x0b \x01(\x08\"\xd1\x01\n\x11PropertiesRequest\x12\x1a\n\x12non_ref_properties\x18\x01 \x03(\t\x12\x39\n\x0eref_properties\x18\x02 \x03(\x0b\x32!.weaviate.v1.RefPropertiesRequest\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\x12$\n\x1creturn_all_nonref_properties\x18\x0b \x01(\x08\"\x8b\x01\n\x17ObjectPropertiesRequest\x12\x11\n\tprop_name\x18\x01 \x01(\t\x12\x1c\n\x14primitive_properties\x18\x02 \x03(\t\x12?\n\x11object_properties\x18\x03 \x03(\x0b\x32$.weaviate.v1.ObjectPropertiesRequest\"\xb1\x01\n\x14RefPropertiesRequest\x12\x1a\n\x12reference_property\x18\x01 \x01(\t\x12\x32\n\nproperties\x18\x02 \x01(\x0b\x32\x1e.weaviate.v1.PropertiesRequest\x12.\n\x08metadata\x18\x03 \x01(\x0b\x32\x1c.weaviate.v1.MetadataRequest\x12\x19\n\x11target_collection\x18\x04 \x01(\t\"8\n\x06Rerank\x12\x10\n\x08property\x18\x01 \x01(\t\x12\x12\n\x05query\x18\x02 \x01(\tH\x00\x88\x01\x01\x42\x08\n\x06_query\"\xf7\x02\n\x0bSearchReply\x12\x0c\n\x04took\x18\x01 \x01(\x02\x12*\n\x07results\x18\x02 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12*\n\x19generative_grouped_result\x18\x03 \x01(\tB\x02\x18\x01H\x00\x88\x01\x01\x12\x34\n\x10group_by_results\x18\x04 \x03(\x0b\x32\x1a.weaviate.v1.GroupByResult\x12\x46\n\x1agenerative_grouped_results\x18\x05 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x01\x88\x01\x01\x12\x35\n\rquery_profile\x18\x06 \x01(\x0b\x32\x19.weaviate.v1.QueryProfileH\x02\x88\x01\x01\x42\x1c\n\x1a_generative_grouped_resultB\x1d\n\x1b_generative_grouped_resultsB\x10\n\x0e_query_profile\"\x9e\x03\n\x0cQueryProfile\x12\x36\n\x06shards\x18\x01 \x03(\x0b\x32&.weaviate.v1.QueryProfile.ShardProfile\x1a\x86\x01\n\rSearchProfile\x12\x45\n\x07\x64\x65tails\x18\x01 \x03(\x0b\x32\x34.weaviate.v1.QueryProfile.SearchProfile.DetailsEntry\x1a.\n\x0c\x44\x65tailsEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01\x1a\xcc\x01\n\x0cShardProfile\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x0c\n\x04node\x18\x02 \x01(\t\x12\x46\n\x08searches\x18\x03 \x03(\x0b\x32\x34.weaviate.v1.QueryProfile.ShardProfile.SearchesEntry\x1aX\n\rSearchesEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\x36\n\x05value\x18\x02 \x01(\x0b\x32\'.weaviate.v1.QueryProfile.SearchProfile:\x02\x38\x01\"\x1c\n\x0bRerankReply\x12\r\n\x05score\x18\x01 \x01(\x01\"\xe9\x02\n\rGroupByResult\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x14\n\x0cmin_distance\x18\x02 \x01(\x02\x12\x14\n\x0cmax_distance\x18\x03 \x01(\x02\x12\x19\n\x11number_of_objects\x18\x04 \x01(\x03\x12*\n\x07objects\x18\x05 \x03(\x0b\x32\x19.weaviate.v1.SearchResult\x12-\n\x06rerank\x18\x06 \x01(\x0b\x32\x18.weaviate.v1.RerankReplyH\x00\x88\x01\x01\x12\x39\n\ngenerative\x18\x07 \x01(\x0b\x32\x1c.weaviate.v1.GenerativeReplyB\x02\x18\x01H\x01\x88\x01\x01\x12=\n\x11generative_result\x18\x08 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x02\x88\x01\x01\x42\t\n\x07_rerankB\r\n\x0b_generativeB\x14\n\x12_generative_result\"\xb7\x01\n\x0cSearchResult\x12\x31\n\nproperties\x18\x01 \x01(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12-\n\x08metadata\x18\x02 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12\x36\n\ngenerative\x18\x03 \x01(\x0b\x32\x1d.weaviate.v1.GenerativeResultH\x00\x88\x01\x01\x42\r\n\x0b_generative\"\xf7\x04\n\x0eMetadataResult\x12\n\n\x02id\x18\x01 \x01(\t\x12\x12\n\x06vector\x18\x02 \x03(\x02\x42\x02\x18\x01\x12\x1a\n\x12\x63reation_time_unix\x18\x03 \x01(\x03\x12\"\n\x1a\x63reation_time_unix_present\x18\x04 \x01(\x08\x12\x1d\n\x15last_update_time_unix\x18\x05 \x01(\x03\x12%\n\x1dlast_update_time_unix_present\x18\x06 \x01(\x08\x12\x10\n\x08\x64istance\x18\x07 \x01(\x02\x12\x18\n\x10\x64istance_present\x18\x08 \x01(\x08\x12\x11\n\tcertainty\x18\t \x01(\x02\x12\x19\n\x11\x63\x65rtainty_present\x18\n \x01(\x08\x12\r\n\x05score\x18\x0b \x01(\x02\x12\x15\n\rscore_present\x18\x0c \x01(\x08\x12\x15\n\rexplain_score\x18\r \x01(\t\x12\x1d\n\x15\x65xplain_score_present\x18\x0e \x01(\x08\x12\x1a\n\ris_consistent\x18\x0f \x01(\x08H\x00\x88\x01\x01\x12\x16\n\ngenerative\x18\x10 \x01(\tB\x02\x18\x01\x12\x1e\n\x12generative_present\x18\x11 \x01(\x08\x42\x02\x18\x01\x12\x1d\n\x15is_consistent_present\x18\x12 \x01(\x08\x12\x14\n\x0cvector_bytes\x18\x13 \x01(\x0c\x12\x13\n\x0bid_as_bytes\x18\x14 \x01(\x0c\x12\x14\n\x0crerank_score\x18\x15 \x01(\x01\x12\x1c\n\x14rerank_score_present\x18\x16 \x01(\x08\x12%\n\x07vectors\x18\x17 \x03(\x0b\x32\x14.weaviate.v1.VectorsB\x10\n\x0e_is_consistent\"\x88\x02\n\x10PropertiesResult\x12\x33\n\tref_props\x18\x02 \x03(\x0b\x32 .weaviate.v1.RefPropertiesResult\x12\x19\n\x11target_collection\x18\x03 \x01(\t\x12-\n\x08metadata\x18\x04 \x01(\x0b\x32\x1b.weaviate.v1.MetadataResult\x12.\n\rnon_ref_props\x18\x0b \x01(\x0b\x32\x17.weaviate.v1.Properties\x12\x1b\n\x13ref_props_requested\x18\x0c \x01(\x08J\x04\x08\x01\x10\x02J\x04\x08\x05\x10\x06J\x04\x08\x06\x10\x07J\x04\x08\x07\x10\x08J\x04\x08\x08\x10\tJ\x04\x08\t\x10\nJ\x04\x08\n\x10\x0b\"[\n\x13RefPropertiesResult\x12\x31\n\nproperties\x18\x01 \x03(\x0b\x32\x1d.weaviate.v1.PropertiesResult\x12\x11\n\tprop_name\x18\x02 \x01(\tBs\n#io.weaviate.client.grpc.protocol.v1B\x16WeaviateProtoSearchGetZ4github.com/weaviate/weaviate/grpc/generated;protocolb\x06proto3')
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -42,6 +42,10 @@
   _globals['_SEARCHREQUEST'].fields_by_name['uses_125_api']._serialized_options = b'\030\001'
   _globals['_SEARCHREPLY'].fields_by_name['generative_grouped_result']._loaded_options = None
   _globals['_SEARCHREPLY'].fields_by_name['generative_grouped_result']._serialized_options = b'\030\001'
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._loaded_options = None
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._serialized_options = b'8\001'
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._loaded_options = None
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._serialized_options = b'8\001'
   _globals['_GROUPBYRESULT'].fields_by_name['generative']._loaded_options = None
   _globals['_GROUPBYRESULT'].fields_by_name['generative']._serialized_options = b'\030\001'
   _globals['_METADATARESULT'].fields_by_name['vector']._loaded_options = None
@@ -57,27 +61,37 @@
   _globals['_SORTBY']._serialized_start=1632
   _globals['_SORTBY']._serialized_end=1673
   _globals['_METADATAREQUEST']._serialized_start=1676
-  _globals['_METADATAREQUEST']._serialized_end=1897
-  _globals['_PROPERTIESREQUEST']._serialized_start=1900
-  _globals['_PROPERTIESREQUEST']._serialized_end=2109
-  _globals['_OBJECTPROPERTIESREQUEST']._serialized_start=2112
-  _globals['_OBJECTPROPERTIESREQUEST']._serialized_end=2251
-  _globals['_REFPROPERTIESREQUEST']._serialized_start=2254
-  _globals['_REFPROPERTIESREQUEST']._serialized_end=2431
-  _globals['_RERANK']._serialized_start=2433
-  _globals['_RERANK']._serialized_end=2489
-  _globals['_SEARCHREPLY']._serialized_start=2492
-  _globals['_SEARCHREPLY']._serialized_end=2794
-  _globals['_RERANKREPLY']._serialized_start=2796
-  _globals['_RERANKREPLY']._serialized_end=2824
-  _globals['_GROUPBYRESULT']._serialized_start=2827
-  _globals['_GROUPBYRESULT']._serialized_end=3188
-  _globals['_SEARCHRESULT']._serialized_start=3191
-  _globals['_SEARCHRESULT']._serialized_end=3374
-  _globals['_METADATARESULT']._serialized_start=3377
-  _globals['_METADATARESULT']._serialized_end=4008
-  _globals['_PROPERTIESRESULT']._serialized_start=4011
-  _globals['_PROPERTIESRESULT']._serialized_end=4275
-  _globals['_REFPROPERTIESRESULT']._serialized_start=4277
-  _globals['_REFPROPERTIESRESULT']._serialized_end=4368
+  _globals['_METADATAREQUEST']._serialized_end=1920
+  _globals['_PROPERTIESREQUEST']._serialized_start=1923
+  _globals['_PROPERTIESREQUEST']._serialized_end=2132
+  _globals['_OBJECTPROPERTIESREQUEST']._serialized_start=2135
+  _globals['_OBJECTPROPERTIESREQUEST']._serialized_end=2274
+  _globals['_REFPROPERTIESREQUEST']._serialized_start=2277
+  _globals['_REFPROPERTIESREQUEST']._serialized_end=2454
+  _globals['_RERANK']._serialized_start=2456
+  _globals['_RERANK']._serialized_end=2512
+  _globals['_SEARCHREPLY']._serialized_start=2515
+  _globals['_SEARCHREPLY']._serialized_end=2890
+  _globals['_QUERYPROFILE']._serialized_start=2893
+  _globals['_QUERYPROFILE']._serialized_end=3307
+  _globals['_QUERYPROFILE_SEARCHPROFILE']._serialized_start=2966
+  _globals['_QUERYPROFILE_SEARCHPROFILE']._serialized_end=3100
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._serialized_start=3054
+  _globals['_QUERYPROFILE_SEARCHPROFILE_DETAILSENTRY']._serialized_end=3100
+  _globals['_QUERYPROFILE_SHARDPROFILE']._serialized_start=3103
+  _globals['_QUERYPROFILE_SHARDPROFILE']._serialized_end=3307
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._serialized_start=3219
+  _globals['_QUERYPROFILE_SHARDPROFILE_SEARCHESENTRY']._serialized_end=3307
+  _globals['_RERANKREPLY']._serialized_start=3309
+  _globals['_RERANKREPLY']._serialized_end=3337
+  _globals['_GROUPBYRESULT']._serialized_start=3340
+  _globals['_GROUPBYRESULT']._serialized_end=3701
+  _globals['_SEARCHRESULT']._serialized_start=3704
+  _globals['_SEARCHRESULT']._serialized_end=3887
+  _globals['_METADATARESULT']._serialized_start=3890
+  _globals['_METADATARESULT']._serialized_end=4521
+  _globals['_PROPERTIESRESULT']._serialized_start=4524
+  _globals['_PROPERTIESRESULT']._serialized_end=4788
+  _globals['_REFPROPERTIESRESULT']._serialized_start=4790
+  _globals['_REFPROPERTIESRESULT']._serialized_end=4881
 # @@protoc_insertion_point(module_scope)
diff --git a/weaviate/proto/v1/v6300/v1/search_get_pb2.pyi b/weaviate/proto/v1/v6300/v1/search_get_pb2.pyi
index 8dd3cb881..f631fb396 100644
--- a/weaviate/proto/v1/v6300/v1/search_get_pb2.pyi
+++ b/weaviate/proto/v1/v6300/v1/search_get_pb2.pyi
@@ -89,7 +89,7 @@ class SortBy(_message.Message):
     def __init__(self, ascending: bool = ..., path: _Optional[_Iterable[str]] = ...) -> None: ...
 
 class MetadataRequest(_message.Message):
-    __slots__ = ("uuid", "vector", "creation_time_unix", "last_update_time_unix", "distance", "certainty", "score", "explain_score", "is_consistent", "vectors")
+    __slots__ = ("uuid", "vector", "creation_time_unix", "last_update_time_unix", "distance", "certainty", "score", "explain_score", "is_consistent", "vectors", "query_profile")
     UUID_FIELD_NUMBER: _ClassVar[int]
     VECTOR_FIELD_NUMBER: _ClassVar[int]
     CREATION_TIME_UNIX_FIELD_NUMBER: _ClassVar[int]
@@ -100,6 +100,7 @@ class MetadataRequest(_message.Message):
     EXPLAIN_SCORE_FIELD_NUMBER: _ClassVar[int]
     IS_CONSISTENT_FIELD_NUMBER: _ClassVar[int]
     VECTORS_FIELD_NUMBER: _ClassVar[int]
+    QUERY_PROFILE_FIELD_NUMBER: _ClassVar[int]
     uuid: bool
     vector: bool
     creation_time_unix: bool
@@ -110,7 +111,8 @@ class MetadataRequest(_message.Message):
     explain_score: bool
     is_consistent: bool
     vectors: _containers.RepeatedScalarFieldContainer[str]
-    def __init__(self, uuid: bool = ..., vector: bool = ..., creation_time_unix: bool = ..., last_update_time_unix: bool = ..., distance: bool = ..., certainty: bool = ..., score: bool = ..., explain_score: bool = ..., is_consistent: bool = ..., vectors: _Optional[_Iterable[str]] = ...) -> None: ...
+    query_profile: bool
+    def __init__(self, uuid: bool = ..., vector: bool = ..., creation_time_unix: bool = ..., last_update_time_unix: bool = ..., distance: bool = ..., certainty: bool = ..., score: bool = ..., explain_score: bool = ..., is_consistent: bool = ..., vectors: _Optional[_Iterable[str]] = ..., query_profile: bool = ...) -> None: ...
 
 class PropertiesRequest(_message.Message):
     __slots__ = ("non_ref_properties", "ref_properties", "object_properties", "return_all_nonref_properties")
@@ -155,18 +157,54 @@ class Rerank(_message.Message):
     def __init__(self, property: _Optional[str] = ..., query: _Optional[str] = ...) -> None: ...
 
 class SearchReply(_message.Message):
-    __slots__ = ("took", "results", "generative_grouped_result", "group_by_results", "generative_grouped_results")
+    __slots__ = ("took", "results", "generative_grouped_result", "group_by_results", "generative_grouped_results", "query_profile")
     TOOK_FIELD_NUMBER: _ClassVar[int]
     RESULTS_FIELD_NUMBER: _ClassVar[int]
     GENERATIVE_GROUPED_RESULT_FIELD_NUMBER: _ClassVar[int]
     GROUP_BY_RESULTS_FIELD_NUMBER: _ClassVar[int]
     GENERATIVE_GROUPED_RESULTS_FIELD_NUMBER: _ClassVar[int]
+    QUERY_PROFILE_FIELD_NUMBER: _ClassVar[int]
     took: float
     results: _containers.RepeatedCompositeFieldContainer[SearchResult]
     generative_grouped_result: str
     group_by_results: _containers.RepeatedCompositeFieldContainer[GroupByResult]
     generative_grouped_results: _generative_pb2.GenerativeResult
-    def __init__(self, took: _Optional[float] = ..., results: _Optional[_Iterable[_Union[SearchResult, _Mapping]]] = ..., generative_grouped_result: _Optional[str] = ..., group_by_results: _Optional[_Iterable[_Union[GroupByResult, _Mapping]]] = ..., generative_grouped_results: _Optional[_Union[_generative_pb2.GenerativeResult, _Mapping]] = ...) -> None: ...
+    query_profile: QueryProfile
+    def __init__(self, took: _Optional[float] = ..., results: _Optional[_Iterable[_Union[SearchResult, _Mapping]]] = ..., generative_grouped_result: _Optional[str] = ..., group_by_results: _Optional[_Iterable[_Union[GroupByResult, _Mapping]]] = ..., generative_grouped_results: _Optional[_Union[_generative_pb2.GenerativeResult, _Mapping]] = ..., query_profile: _Optional[_Union[QueryProfile, _Mapping]] = ...) -> None: ...
+
+class QueryProfile(_message.Message):
+    __slots__ = ("shards",)
+    class SearchProfile(_message.Message):
+        __slots__ = ("details",)
+        class DetailsEntry(_message.Message):
+            __slots__ = ("key", "value")
+            KEY_FIELD_NUMBER: _ClassVar[int]
+            VALUE_FIELD_NUMBER: _ClassVar[int]
+            key: str
+            value: str
+            def __init__(self, key: _Optional[str] = ..., value: _Optional[str] = ...) -> None: ...
+        DETAILS_FIELD_NUMBER: _ClassVar[int]
+        details: _containers.ScalarMap[str, str]
+        def __init__(self, details: _Optional[_Mapping[str, str]] = ...) -> None: ...
+    class ShardProfile(_message.Message):
+        __slots__ = ("name", "node", "searches")
+        class SearchesEntry(_message.Message):
+            __slots__ = ("key", "value")
+            KEY_FIELD_NUMBER: _ClassVar[int]
+            VALUE_FIELD_NUMBER: _ClassVar[int]
+            key: str
+            value: QueryProfile.SearchProfile
+            def __init__(self, key: _Optional[str] = ..., value: _Optional[_Union[QueryProfile.SearchProfile, _Mapping]] = ...) -> None: ...
+        NAME_FIELD_NUMBER: _ClassVar[int]
+        NODE_FIELD_NUMBER: _ClassVar[int]
+        SEARCHES_FIELD_NUMBER: _ClassVar[int]
+        name: str
+        node: str
+        searches: _containers.MessageMap[str, QueryProfile.SearchProfile]
+        def __init__(self, name: _Optional[str] = ..., node: _Optional[str] = ..., searches: _Optional[_Mapping[str, QueryProfile.SearchProfile]] = ...) -> None: ...
+    SHARDS_FIELD_NUMBER: _ClassVar[int]
+    shards: _containers.RepeatedCompositeFieldContainer[QueryProfile.ShardProfile]
+    def __init__(self, shards: _Optional[_Iterable[_Union[QueryProfile.ShardProfile, _Mapping]]] = ...) -> None: ...
 
 class RerankReply(_message.Message):
     __slots__ = ("score",)

From 6fd60b567010b19399e4e1accfe8261103f6c989 Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Tue, 14 Apr 2026 16:05:53 +0200
Subject: [PATCH 55/99] Reformatted

---
 integration/test_collection_query_profile.py | 14 +++++---------
 1 file changed, 5 insertions(+), 9 deletions(-)

diff --git a/integration/test_collection_query_profile.py b/integration/test_collection_query_profile.py
index 021c83802..2061eea7a 100644
--- a/integration/test_collection_query_profile.py
+++ b/integration/test_collection_query_profile.py
@@ -13,9 +13,9 @@
 
 def assert_go_duration(value: str, label: str = "") -> None:
     """Assert that a string looks like a Go duration (e.g. '1.234ms', '5.458µs')."""
-    assert GO_DURATION_RE.fullmatch(
-        value
-    ), f"Expected Go duration format for {label!r}, got {value!r}"
+    assert GO_DURATION_RE.fullmatch(value), (
+        f"Expected Go duration format for {label!r}, got {value!r}"
+    )
 
 
 def assert_common_profile(profile: SearchProfileReturn) -> None:
@@ -91,16 +91,12 @@ def test_near_vector_with_query_profile(collection_with_data):
     assert_common_profile(vector_profile)
 
     assert "vector_search_took" in vector_profile.details
-    assert_go_duration(
-        vector_profile.details["vector_search_took"], "vector_search_took"
-    )
+    assert_go_duration(vector_profile.details["vector_search_took"], "vector_search_took")
 
     assert "hnsw_flat_search" in vector_profile.details
     assert vector_profile.details["hnsw_flat_search"] in ("true", "false")
 
-    layer_keys = [
-        k for k in vector_profile.details if k.startswith("knn_search_layer_")
-    ]
+    layer_keys = [k for k in vector_profile.details if k.startswith("knn_search_layer_")]
     assert len(layer_keys) > 0, "Expected at least one knn_search_layer_*_took key"
     for k in layer_keys:
         assert_go_duration(vector_profile.details[k], k)

From a1df0988a5094b1f97deaae8a0a59dbd1b139760 Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Tue, 14 Apr 2026 16:23:13 +0200
Subject: [PATCH 56/99] Skip test for lower versions

---
 integration/test_collection_query_profile.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/integration/test_collection_query_profile.py b/integration/test_collection_query_profile.py
index 2061eea7a..92687129d 100644
--- a/integration/test_collection_query_profile.py
+++ b/integration/test_collection_query_profile.py
@@ -37,6 +37,8 @@ def client():
 
 @pytest.fixture(scope="module")
 def collection_with_data(client: weaviate.WeaviateClient):
+    if client._connection._weaviate_version.is_lower_than(1, 36, 9):
+        pytest.skip("Query profiling requires Weaviate >= 1.36.9")
     name = "TestQueryProfile"
     client.collections.delete(name)
     collection = client.collections.create(

From 239ed323ccc92dcd1713a5c0b372b7036dd70a79 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 14 Apr 2026 21:03:13 +0100
Subject: [PATCH 57/99] feat: add tokenizer module with sync and async support,
 including integration tests

---
 integration/test_tokenize.py  | 355 ++++++++++++++++++++++++++++++++++
 weaviate/__init__.py          |   2 +
 weaviate/client.py            |   3 +
 weaviate/client.pyi           |   3 +
 weaviate/tokenize/__init__.py |   7 +
 weaviate/tokenize/async_.py   |   8 +
 weaviate/tokenize/executor.py | 166 ++++++++++++++++
 weaviate/tokenize/sync.py     |   8 +
 weaviate/tokenize/types.py    |  25 +++
 9 files changed, 577 insertions(+)
 create mode 100644 integration/test_tokenize.py
 create mode 100644 weaviate/tokenize/__init__.py
 create mode 100644 weaviate/tokenize/async_.py
 create mode 100644 weaviate/tokenize/executor.py
 create mode 100644 weaviate/tokenize/sync.py
 create mode 100644 weaviate/tokenize/types.py

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
new file mode 100644
index 000000000..e54f9d49d
--- /dev/null
+++ b/integration/test_tokenize.py
@@ -0,0 +1,355 @@
+"""Integration tests for the tokenize module.
+
+These tests cover the client's responsibilities:
+- Correct serialization of inputs (enums, _TextAnalyzerConfigCreate, kwargs)
+- Correct deserialization of responses into typed objects
+- Client-side validation (_TextAnalyzerConfigCreate rejects invalid input)
+- Both sync and async client paths
+"""
+
+from typing import AsyncGenerator, Generator
+
+import pytest
+import pytest_asyncio
+
+import weaviate
+from weaviate.collections.classes.config import (
+    StopwordsConfig,
+    StopwordsPreset,
+    TextAnalyzerConfig,
+    Tokenization,
+    _StopwordsCreate,
+    _TextAnalyzerConfigCreate,
+)
+from weaviate.config import AdditionalConfig
+from weaviate.tokenize.types import TokenizeResult
+
+
+@pytest.fixture(scope="module")
+def client() -> Generator[weaviate.WeaviateClient, None, None]:
+    c = weaviate.connect_to_local(
+        additional_config=AdditionalConfig(timeout=(60, 120)),
+    )
+    yield c
+    c.close()
+
+
+@pytest_asyncio.fixture
+async def async_client() -> AsyncGenerator[weaviate.WeaviateAsyncClient, None]:
+    c = weaviate.use_async_with_local(
+        additional_config=AdditionalConfig(timeout=(60, 120)),
+    )
+    await c.connect()
+    yield c
+    await c.close()
+
+
+# ---------------------------------------------------------------------------
+# Serialization: enums, strings, kwargs, _TextAnalyzerConfigCreate
+# ---------------------------------------------------------------------------
+
+
+class TestSerialization:
+    """Verify the client correctly serializes different input forms."""
+
+    @pytest.mark.parametrize(
+        "tokenization,text,expected_tokens",
+        [
+            (Tokenization.WORD, "The quick brown fox", ["the", "quick", "brown", "fox"]),
+            (Tokenization.LOWERCASE, "Hello World Test", ["hello", "world", "test"]),
+            (Tokenization.WHITESPACE, "Hello World Test", ["Hello", "World", "Test"]),
+            (Tokenization.FIELD, "  Hello World  ", ["Hello World"]),
+            (Tokenization.TRIGRAM, "Hello", ["hel", "ell", "llo"]),
+        ],
+    )
+    def test_tokenization_enum(
+        self,
+        client: weaviate.WeaviateClient,
+        tokenization: Tokenization,
+        text: str,
+        expected_tokens: list,
+    ) -> None:
+        result = client.tokenize.text(text=text, tokenization=tokenization)
+        assert isinstance(result, TokenizeResult)
+        assert result.tokenization == tokenization.value
+        assert result.indexed == expected_tokens
+        assert result.query == expected_tokens
+
+    def test_tokenization_string(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenize.text(text="hello world", tokenization="word")
+        assert result.tokenization == "word"
+        assert result.indexed == ["hello", "world"]
+
+    def test_stopword_preset_enum(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenize.text(
+            text="The quick brown fox",
+            tokenization=Tokenization.WORD,
+            stopword_preset=StopwordsPreset.EN,
+        )
+        assert "the" not in result.query
+        assert "quick" in result.query
+
+    def test_stopword_preset_string(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenize.text(
+            text="The quick brown fox",
+            tokenization=Tokenization.WORD,
+            stopword_preset="en",
+        )
+        assert "the" not in result.query
+
+    def test_ascii_fold_via_kwargs(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenize.text(
+            text="L'école est fermée",
+            tokenization=Tokenization.WORD,
+            ascii_fold=True,
+        )
+        assert result.indexed == ["l", "ecole", "est", "fermee"]
+
+    def test_ascii_fold_via_analyzer_config(self, client: weaviate.WeaviateClient) -> None:
+        cfg = _TextAnalyzerConfigCreate(ascii_fold=True)
+        result = client.tokenize.text(
+            text="L'école est fermée",
+            tokenization=Tokenization.WORD,
+            analyzer_config=cfg,
+        )
+        assert result.indexed == ["l", "ecole", "est", "fermee"]
+
+    def test_analyzer_config_and_kwargs_produce_same_result(
+        self, client: weaviate.WeaviateClient
+    ) -> None:
+        """analyzer_config object and equivalent kwargs must produce identical output."""
+        cfg = _TextAnalyzerConfigCreate(
+            ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset=StopwordsPreset.EN
+        )
+        via_config = client.tokenize.text(
+            text="L'école est fermée",
+            tokenization=Tokenization.WORD,
+            analyzer_config=cfg,
+        )
+        via_kwargs = client.tokenize.text(
+            text="L'école est fermée",
+            tokenization=Tokenization.WORD,
+            ascii_fold=True,
+            ascii_fold_ignore=["é"],
+            stopword_preset=StopwordsPreset.EN,
+        )
+        assert via_config.indexed == via_kwargs.indexed
+        assert via_config.query == via_kwargs.query
+
+    def test_stopword_presets_serialization(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenize.text(
+            text="hello world test",
+            tokenization=Tokenization.WORD,
+            stopword_preset="custom",
+            stopword_presets={
+                "custom": _StopwordsCreate(preset=None, additions=["test"], removals=None),
+            },
+        )
+        assert result.indexed == ["hello", "world", "test"]
+        assert result.query == ["hello", "world"]
+
+    def test_stopword_presets_with_base_and_removals(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenize.text(
+            text="the quick",
+            tokenization=Tokenization.WORD,
+            stopword_preset="en-no-the",
+            stopword_presets={
+                "en-no-the": _StopwordsCreate(
+                    preset=StopwordsPreset.EN, additions=None, removals=["the"]
+                ),
+            },
+        )
+        assert result.indexed == ["the", "quick"]
+        assert result.query == ["the", "quick"]
+
+
+# ---------------------------------------------------------------------------
+# Deserialization: typed response fields
+# ---------------------------------------------------------------------------
+
+
+class TestDeserialization:
+    """Verify the client correctly deserializes response fields into typed objects."""
+
+    def test_result_type(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenize.text(text="hello", tokenization=Tokenization.WORD)
+        assert isinstance(result, TokenizeResult)
+        assert isinstance(result.indexed, list)
+        assert isinstance(result.query, list)
+
+    def test_analyzer_config_deserialized(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenize.text(
+            text="L'école",
+            tokenization=Tokenization.WORD,
+            ascii_fold=True,
+            ascii_fold_ignore=["é"],
+            stopword_preset=StopwordsPreset.EN,
+        )
+        assert isinstance(result.analyzer_config, TextAnalyzerConfig)
+        assert result.analyzer_config.ascii_fold is True
+        assert result.analyzer_config.ascii_fold_ignore == ["é"]
+        assert result.analyzer_config.stopword_preset == "en"
+
+    def test_no_analyzer_config_returns_none(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenize.text(text="hello", tokenization=Tokenization.WORD)
+        assert result.analyzer_config is None
+
+    def test_stopword_config_deserialized_on_property(
+        self, client: weaviate.WeaviateClient
+    ) -> None:
+        """Property endpoint returns stopwordConfig; verify it deserializes to StopwordsConfig."""
+        client.collections.delete("TestDeserStopword")
+        try:
+            client.collections.create_from_dict(
+                {
+                    "class": "TestDeserStopword",
+                    "vectorizer": "none",
+                    "properties": [
+                        {
+                            "name": "title",
+                            "dataType": ["text"],
+                            "tokenization": "word",
+                            "textAnalyzer": {"stopwordPreset": "en"},
+                        },
+                    ],
+                }
+            )
+            result = client.tokenize.property(
+                collection_name="TestDeserStopword",
+                property_name="title",
+                text="the quick",
+            )
+            assert isinstance(result, TokenizeResult)
+            assert result.tokenization == "word"
+            # Stopword config should be deserialized when present
+            if result.stopword_config is not None:
+                assert isinstance(result.stopword_config, StopwordsConfig)
+        finally:
+            client.collections.delete("TestDeserStopword")
+
+    def test_property_result_types(self, client: weaviate.WeaviateClient) -> None:
+        client.collections.delete("TestDeserPropTypes")
+        try:
+            client.collections.create_from_dict(
+                {
+                    "class": "TestDeserPropTypes",
+                    "vectorizer": "none",
+                    "properties": [
+                        {
+                            "name": "tag",
+                            "dataType": ["text"],
+                            "tokenization": "field",
+                        },
+                    ],
+                }
+            )
+            result = client.tokenize.property(
+                collection_name="TestDeserPropTypes",
+                property_name="tag",
+                text="  Hello World  ",
+            )
+            assert isinstance(result, TokenizeResult)
+            assert result.tokenization == "field"
+            assert result.indexed == ["Hello World"]
+        finally:
+            client.collections.delete("TestDeserPropTypes")
+
+
+# ---------------------------------------------------------------------------
+# Client-side validation (_TextAnalyzerConfigCreate)
+# ---------------------------------------------------------------------------
+
+
+class TestClientSideValidation:
+    """Verify that _TextAnalyzerConfigCreate rejects invalid input before hitting the server."""
+
+    def test_ascii_fold_ignore_without_fold_raises(self) -> None:
+        with pytest.raises(ValueError, match="asciiFoldIgnore"):
+            _TextAnalyzerConfigCreate(ascii_fold=False, ascii_fold_ignore=["é"])
+
+    def test_ascii_fold_ignore_without_fold_default_raises(self) -> None:
+        with pytest.raises(ValueError, match="asciiFoldIgnore"):
+            _TextAnalyzerConfigCreate(ascii_fold_ignore=["é"])
+
+    def test_valid_config_does_not_raise(self) -> None:
+        cfg = _TextAnalyzerConfigCreate(ascii_fold=True, ascii_fold_ignore=["é", "ñ"])
+        assert cfg.asciiFold is True
+        assert cfg.asciiFoldIgnore == ["é", "ñ"]
+
+    def test_fold_without_ignore_is_valid(self) -> None:
+        cfg = _TextAnalyzerConfigCreate(ascii_fold=True)
+        assert cfg.asciiFold is True
+        assert cfg.asciiFoldIgnore is None
+
+    def test_stopword_preset_only_is_valid(self) -> None:
+        cfg = _TextAnalyzerConfigCreate(stopword_preset="en")
+        assert cfg.stopwordPreset == "en"
+
+    def test_empty_config_is_valid(self) -> None:
+        cfg = _TextAnalyzerConfigCreate()
+        assert cfg.asciiFold is None
+        assert cfg.asciiFoldIgnore is None
+        assert cfg.stopwordPreset is None
+
+
+# ---------------------------------------------------------------------------
+# Async client
+# ---------------------------------------------------------------------------
+
+
+class TestAsyncClient:
+    """Verify both text() and property() work through the async client."""
+
+    @pytest.mark.asyncio
+    async def test_text_tokenize(self, async_client: weaviate.WeaviateAsyncClient) -> None:
+        result = await async_client.tokenize.text(
+            text="The quick brown fox",
+            tokenization=Tokenization.WORD,
+        )
+        assert isinstance(result, TokenizeResult)
+        assert result.indexed == ["the", "quick", "brown", "fox"]
+
+    @pytest.mark.asyncio
+    async def test_text_with_analyzer_config(
+        self, async_client: weaviate.WeaviateAsyncClient
+    ) -> None:
+        cfg = _TextAnalyzerConfigCreate(ascii_fold=True, stopword_preset=StopwordsPreset.EN)
+        result = await async_client.tokenize.text(
+            text="L'école est fermée",
+            tokenization=Tokenization.WORD,
+            analyzer_config=cfg,
+        )
+        assert result.indexed == ["l", "ecole", "est", "fermee"]
+        assert isinstance(result.analyzer_config, TextAnalyzerConfig)
+        assert result.analyzer_config.ascii_fold is True
+
+    @pytest.mark.asyncio
+    async def test_property_tokenize(self, async_client: weaviate.WeaviateAsyncClient) -> None:
+        await async_client.collections.delete("TestAsyncPropTokenize")
+        try:
+            await async_client.collections.create_from_dict(
+                {
+                    "class": "TestAsyncPropTokenize",
+                    "vectorizer": "none",
+                    "properties": [
+                        {
+                            "name": "title",
+                            "dataType": ["text"],
+                            "tokenization": "word",
+                            "textAnalyzer": {"stopwordPreset": "en"},
+                        },
+                    ],
+                }
+            )
+            result = await async_client.tokenize.property(
+                collection_name="TestAsyncPropTokenize",
+                property_name="title",
+                text="The quick brown fox",
+            )
+            assert isinstance(result, TokenizeResult)
+            assert result.tokenization == "word"
+            assert result.indexed == ["the", "quick", "brown", "fox"]
+            assert "the" not in result.query
+            assert "quick" in result.query
+        finally:
+            await async_client.collections.delete("TestAsyncPropTokenize")
diff --git a/weaviate/__init__.py b/weaviate/__init__.py
index 562b142bc..6fd9368ea 100644
--- a/weaviate/__init__.py
+++ b/weaviate/__init__.py
@@ -21,6 +21,7 @@
     embedded,
     exceptions,
     outputs,
+    tokenize,
     types,
 )
 from .client import Client, WeaviateAsyncClient, WeaviateClient
@@ -67,6 +68,7 @@
     "embedded",
     "exceptions",
     "outputs",
+    "tokenize",
     "types",
     "use_async_with_custom",
     "use_async_with_embedded",
diff --git a/weaviate/client.py b/weaviate/client.py
index d7f9080f4..cbd12be9a 100644
--- a/weaviate/client.py
+++ b/weaviate/client.py
@@ -22,6 +22,7 @@
 from .embedded import EmbeddedOptions
 from .groups import _Groups, _GroupsAsync
 from .rbac import _Roles, _RolesAsync
+from .tokenize import _Tokenize, _TokenizeAsync
 from .types import NUMBER
 from .users import _Users, _UsersAsync
 
@@ -82,6 +83,7 @@ def __init__(
         self.debug = _DebugAsync(self._connection)
         self.groups = _GroupsAsync(self._connection)
         self.roles = _RolesAsync(self._connection)
+        self.tokenize = _TokenizeAsync(self._connection)
         self.users = _UsersAsync(self._connection)
 
     async def __aenter__(self) -> "WeaviateAsyncClient":
@@ -157,6 +159,7 @@ def __init__(
         self.debug = _Debug(self._connection)
         self.groups = _Groups(self._connection)
         self.roles = _Roles(self._connection)
+        self.tokenize = _Tokenize(self._connection)
         self.users = _Users(self._connection)
 
     def __enter__(self) -> "WeaviateClient":
diff --git a/weaviate/client.pyi b/weaviate/client.pyi
index 9b32af15f..a6a44f8f7 100644
--- a/weaviate/client.pyi
+++ b/weaviate/client.pyi
@@ -21,6 +21,7 @@ from .cluster import _Cluster, _ClusterAsync
 from .collections.batch.client import _BatchClientWrapper, _BatchClientWrapperAsync
 from .debug import _Debug, _DebugAsync
 from .rbac import _Roles, _RolesAsync
+from .tokenize import _Tokenize, _TokenizeAsync
 from .types import NUMBER
 
 TIMEOUT_TYPE = Union[Tuple[NUMBER, NUMBER], NUMBER]
@@ -35,6 +36,7 @@ class WeaviateAsyncClient(_WeaviateClientExecutor[ConnectionAsync]):
     debug: _DebugAsync
     groups: _GroupsAsync
     roles: _RolesAsync
+    tokenize: _TokenizeAsync
     users: _UsersAsync
 
     async def close(self) -> None: ...
@@ -58,6 +60,7 @@ class WeaviateClient(_WeaviateClientExecutor[ConnectionSync]):
     debug: _Debug
     groups: _Groups
     roles: _Roles
+    tokenize: _Tokenize
     users: _Users
 
     def close(self) -> None: ...
diff --git a/weaviate/tokenize/__init__.py b/weaviate/tokenize/__init__.py
new file mode 100644
index 000000000..d0c2883c5
--- /dev/null
+++ b/weaviate/tokenize/__init__.py
@@ -0,0 +1,7 @@
+"""Module for tokenize operations."""
+
+from .async_ import _TokenizeAsync
+from .sync import _Tokenize
+from .types import TokenizeResult
+
+__all__ = ["_Tokenize", "_TokenizeAsync", "TokenizeResult"]
diff --git a/weaviate/tokenize/async_.py b/weaviate/tokenize/async_.py
new file mode 100644
index 000000000..a59c392ea
--- /dev/null
+++ b/weaviate/tokenize/async_.py
@@ -0,0 +1,8 @@
+from weaviate.connect import executor
+from weaviate.connect.v4 import ConnectionAsync
+from weaviate.tokenize.executor import _TokenizeExecutor
+
+
+@executor.wrap("async")
+class _TokenizeAsync(_TokenizeExecutor[ConnectionAsync]):
+    pass
diff --git a/weaviate/tokenize/executor.py b/weaviate/tokenize/executor.py
new file mode 100644
index 000000000..bd2c24dc1
--- /dev/null
+++ b/weaviate/tokenize/executor.py
@@ -0,0 +1,166 @@
+"""Tokenize executor."""
+
+from typing import Any, Dict, Generic, List, Optional, Union
+
+from httpx import Response
+
+from weaviate.collections.classes.config import (
+    StopwordsConfig,
+    StopwordsPreset,
+    TextAnalyzerConfig,
+    Tokenization,
+    _StopwordsCreate,
+    _TextAnalyzerConfigCreate,
+)
+from weaviate.connect import executor
+from weaviate.connect.v4 import ConnectionType, _ExpectedStatusCodes
+from weaviate.tokenize.types import TokenizeResult
+
+
+def _parse_analyzer_config(body: Dict[str, Any]) -> Optional[TextAnalyzerConfig]:
+    ac = body.get("analyzerConfig")
+    if ac is None:
+        return None
+    if "asciiFold" not in ac and "stopwordPreset" not in ac:
+        return None
+    return TextAnalyzerConfig(
+        ascii_fold=ac.get("asciiFold", False),
+        ascii_fold_ignore=ac.get("asciiFoldIgnore"),
+        stopword_preset=ac.get("stopwordPreset"),
+    )
+
+
+def _parse_stopword_config(body: Dict[str, Any]) -> Optional[StopwordsConfig]:
+    sc = body.get("stopwordConfig")
+    if sc is None:
+        return None
+    return StopwordsConfig(
+        preset=StopwordsPreset(sc["preset"]) if sc.get("preset") else StopwordsPreset.NONE,
+        additions=sc.get("additions"),
+        removals=sc.get("removals"),
+    )
+
+
+def _parse_tokenize_result(body: Dict[str, Any]) -> TokenizeResult:
+    return TokenizeResult(
+        tokenization=body["tokenization"],
+        indexed=body["indexed"],
+        query=body["query"],
+        analyzer_config=_parse_analyzer_config(body),
+        stopword_config=_parse_stopword_config(body),
+    )
+
+
+class _TokenizeExecutor(Generic[ConnectionType]):
+    def __init__(self, connection: ConnectionType):
+        self._connection = connection
+
+    def text(
+        self,
+        text: str,
+        tokenization: Union[Tokenization, str],
+        *,
+        analyzer_config: Optional[_TextAnalyzerConfigCreate] = None,
+        ascii_fold: Optional[bool] = None,
+        ascii_fold_ignore: Optional[List[str]] = None,
+        stopword_preset: Optional[Union[StopwordsPreset, str]] = None,
+        stopword_presets: Optional[Dict[str, _StopwordsCreate]] = None,
+    ) -> executor.Result[TokenizeResult]:
+        """Tokenize text using the generic /v1/tokenize endpoint.
+
+        Analyzer settings can be provided either via a ``_TextAnalyzerConfigCreate``
+        object **or** via the individual keyword arguments (``ascii_fold``,
+        ``ascii_fold_ignore``, ``stopword_preset``).  If ``analyzer_config`` is
+        given the individual keyword arguments are ignored.
+
+        Args:
+            text: The text to tokenize.
+            tokenization: The tokenization method to use (e.g. Tokenization.WORD).
+            analyzer_config: A ``_TextAnalyzerConfigCreate`` instance that bundles
+                ascii_fold, ascii_fold_ignore, and stopword_preset settings.
+            ascii_fold: Whether to fold accented characters to ASCII equivalents.
+            ascii_fold_ignore: Characters to exclude from ASCII folding.
+            stopword_preset: Stopword preset name to apply for query-time filtering.
+            stopword_presets: Custom stopword preset definitions, keyed by name.
+                Each value is a ``_StopwordsCreate`` with optional preset, additions,
+                and removals fields.
+
+        Returns:
+            A TokenizeResult with indexed and query token lists.
+        """
+        tokenization_str = (
+            tokenization.value if isinstance(tokenization, Tokenization) else tokenization
+        )
+
+        payload: Dict[str, Any] = {
+            "text": text,
+            "tokenization": tokenization_str,
+        }
+
+        if analyzer_config is not None:
+            ac_dict = analyzer_config._to_dict()
+            if ac_dict:
+                payload["analyzerConfig"] = ac_dict
+        else:
+            ac: Dict[str, Any] = {}
+            if ascii_fold is not None:
+                ac["asciiFold"] = ascii_fold
+            if ascii_fold_ignore is not None:
+                ac["asciiFoldIgnore"] = ascii_fold_ignore
+            if stopword_preset is not None:
+                ac["stopwordPreset"] = (
+                    stopword_preset.value
+                    if isinstance(stopword_preset, StopwordsPreset)
+                    else stopword_preset
+                )
+            if ac:
+                payload["analyzerConfig"] = ac
+
+        if stopword_presets is not None:
+            payload["stopwordPresets"] = {
+                name: cfg._to_dict() for name, cfg in stopword_presets.items()
+            }
+
+        def resp(response: Response) -> TokenizeResult:
+            return _parse_tokenize_result(response.json())
+
+        return executor.execute(
+            response_callback=resp,
+            method=self._connection.post,
+            path="/tokenize",
+            weaviate_object=payload,
+            error_msg="Tokenization failed",
+            status_codes=_ExpectedStatusCodes(ok_in=[200], error="tokenize text"),
+        )
+
+    def property(
+        self,
+        collection_name: str,
+        property_name: str,
+        text: str,
+    ) -> executor.Result[TokenizeResult]:
+        """Tokenize text using a property's configured tokenization settings.
+
+        Args:
+            collection_name: The collection (class) name.
+            property_name: The property name whose tokenization config to use.
+            text: The text to tokenize.
+
+        Returns:
+            A TokenizeResult with indexed and query token lists.
+        """
+        path = f"/schema/{collection_name}/properties/{property_name}/tokenize"
+
+        payload: Dict[str, Any] = {"text": text}
+
+        def resp(response: Response) -> TokenizeResult:
+            return _parse_tokenize_result(response.json())
+
+        return executor.execute(
+            response_callback=resp,
+            method=self._connection.post,
+            path=path,
+            weaviate_object=payload,
+            error_msg="Property tokenization failed",
+            status_codes=_ExpectedStatusCodes(ok_in=[200], error="tokenize property text"),
+        )
diff --git a/weaviate/tokenize/sync.py b/weaviate/tokenize/sync.py
new file mode 100644
index 000000000..755c42559
--- /dev/null
+++ b/weaviate/tokenize/sync.py
@@ -0,0 +1,8 @@
+from weaviate.connect import executor
+from weaviate.connect.v4 import ConnectionSync
+from weaviate.tokenize.executor import _TokenizeExecutor
+
+
+@executor.wrap("sync")
+class _Tokenize(_TokenizeExecutor[ConnectionSync]):
+    pass
diff --git a/weaviate/tokenize/types.py b/weaviate/tokenize/types.py
new file mode 100644
index 000000000..ba4009b2d
--- /dev/null
+++ b/weaviate/tokenize/types.py
@@ -0,0 +1,25 @@
+"""Return types for tokenize operations."""
+
+from dataclasses import dataclass, field
+from typing import List, Optional
+
+from weaviate.collections.classes.config import StopwordsConfig, TextAnalyzerConfig
+
+
+@dataclass
+class TokenizeResult:
+    """Result of a tokenization operation.
+
+    Attributes:
+        tokenization: The tokenization method that was applied.
+        indexed: Tokens as they would be stored in the inverted index.
+        query: Tokens as they would be used for querying (after stopword removal).
+        analyzer_config: The text analyzer configuration that was used, if any.
+        stopword_config: The stopword configuration that was used, if any.
+    """
+
+    tokenization: str
+    indexed: List[str]
+    query: List[str]
+    analyzer_config: Optional[TextAnalyzerConfig] = field(default=None)
+    stopword_config: Optional[StopwordsConfig] = field(default=None)

From 480dbe0d4e614606dba54bbe2064af9401fc0f71 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Mon, 9 Mar 2026 16:49:40 +0100
Subject: [PATCH 58/99] Add support for collection export endpoint

---
 integration/test_export.py   | 219 +++++++++++++++++++++++
 weaviate/classes/__init__.py |   2 +
 weaviate/classes/export.py   |  11 ++
 weaviate/client.py           |   3 +
 weaviate/exceptions.py       |   8 +
 weaviate/export/__init__.py  |   7 +
 weaviate/export/async_.py    |   8 +
 weaviate/export/async_.pyi   |  38 ++++
 weaviate/export/executor.py  | 331 +++++++++++++++++++++++++++++++++++
 weaviate/export/export.py    |  92 ++++++++++
 weaviate/export/sync.py      |   8 +
 weaviate/export/sync.pyi     |  38 ++++
 weaviate/outputs/__init__.py |  15 +-
 weaviate/outputs/export.py   |  15 ++
 14 files changed, 794 insertions(+), 1 deletion(-)
 create mode 100644 integration/test_export.py
 create mode 100644 weaviate/classes/export.py
 create mode 100644 weaviate/export/__init__.py
 create mode 100644 weaviate/export/async_.py
 create mode 100644 weaviate/export/async_.pyi
 create mode 100644 weaviate/export/executor.py
 create mode 100644 weaviate/export/export.py
 create mode 100644 weaviate/export/sync.py
 create mode 100644 weaviate/export/sync.pyi
 create mode 100644 weaviate/outputs/export.py

diff --git a/integration/test_export.py b/integration/test_export.py
new file mode 100644
index 000000000..791360e3a
--- /dev/null
+++ b/integration/test_export.py
@@ -0,0 +1,219 @@
+import time
+import uuid
+from typing import Generator, List, Union
+
+import pytest
+from _pytest.fixtures import SubRequest
+
+import weaviate
+from weaviate.auth import Auth
+from weaviate.collections.classes.config import DataType, Property
+from weaviate.exceptions import UnexpectedStatusCodeException
+from weaviate.export.export import (
+    ExportFileFormat,
+    ExportStatus,
+    ExportStorage,
+)
+
+from .conftest import _sanitize_collection_name
+
+RBAC_PORTS = (8093, 50065)
+RBAC_AUTH_CREDS = Auth.api_key("admin-key")
+
+pytestmark = pytest.mark.xdist_group(name="export")
+
+BACKEND = ExportStorage.FILESYSTEM
+
+COLLECTION_NAME = "ExportTestCollection"
+
+OBJECT_PROPS = [{"title": f"object {i}", "count": i} for i in range(5)]
+
+OBJECT_IDS = [
+    "fd34ccf4-1a2a-47ad-8446-231839366c3f",
+    "2653442b-05d8-4fa3-b46a-d4a152eb63bc",
+    "55374edb-17de-487f-86cb-9a9fbc30823f",
+    "124ff6aa-597f-44d0-8c13-62fbb1e66888",
+    "f787386e-7d1c-481f-b8c3-3dbfd8bbad85",
+]
+
+
+@pytest.fixture(scope="module")
+def client() -> Generator[weaviate.WeaviateClient, None, None]:
+    client = weaviate.connect_to_local(
+        port=RBAC_PORTS[0], grpc_port=RBAC_PORTS[1], auth_credentials=RBAC_AUTH_CREDS
+    )
+    client.collections.delete(COLLECTION_NAME)
+
+    col = client.collections.create(
+        name=COLLECTION_NAME,
+        properties=[
+            Property(name="title", data_type=DataType.TEXT),
+            Property(name="count", data_type=DataType.INT),
+        ],
+    )
+    for i, props in enumerate(OBJECT_PROPS):
+        col.data.insert(properties=props, uuid=OBJECT_IDS[i])
+
+    yield client
+    client.collections.delete(COLLECTION_NAME)
+    client.close()
+
+
+def unique_export_id(name: str) -> str:
+    """Generate a unique export ID based on the test name."""
+    name = _sanitize_collection_name(name)
+    random_part = str(uuid.uuid4()).replace("-", "")[:12]
+    return name + random_part
+
+
+def test_create_export_with_waiting(client: weaviate.WeaviateClient, request: SubRequest) -> None:
+    """Create an export with wait_for_completion=True."""
+    export_id = unique_export_id(request.node.name)
+
+    resp = client.export.create(
+        export_id=export_id,
+        backend=BACKEND,
+        include_collections=[COLLECTION_NAME],
+        wait_for_completion=True,
+    )
+    assert resp.status == ExportStatus.SUCCESS
+    assert COLLECTION_NAME in resp.collections
+
+
+def test_create_export_without_waiting(
+    client: weaviate.WeaviateClient, request: SubRequest
+) -> None:
+    """Create an export without waiting, then poll status."""
+    export_id = unique_export_id(request.node.name)
+
+    resp = client.export.create(
+        export_id=export_id,
+        backend=BACKEND,
+        include_collections=[COLLECTION_NAME],
+    )
+    assert resp.status in [ExportStatus.STARTED, ExportStatus.TRANSFERRING, ExportStatus.SUCCESS]
+
+    # poll until done
+    while True:
+        status = client.export.get_status(export_id=export_id, backend=BACKEND)
+        assert status.status in [
+            ExportStatus.STARTED,
+            ExportStatus.TRANSFERRING,
+            ExportStatus.SUCCESS,
+        ]
+        if status.status == ExportStatus.SUCCESS:
+            break
+        time.sleep(0.1)
+
+    assert status.export_id == export_id
+
+
+def test_get_export_status(client: weaviate.WeaviateClient, request: SubRequest) -> None:
+    """Check status of a completed export."""
+    export_id = unique_export_id(request.node.name)
+
+    client.export.create(
+        export_id=export_id,
+        backend=BACKEND,
+        include_collections=[COLLECTION_NAME],
+        wait_for_completion=True,
+    )
+
+    status = client.export.get_status(export_id=export_id, backend=BACKEND)
+    assert status.status == ExportStatus.SUCCESS
+    assert status.export_id == export_id
+    assert status.backend == BACKEND.value
+
+
+def test_create_export_with_parquet_format(
+    client: weaviate.WeaviateClient, request: SubRequest
+) -> None:
+    """Create an export explicitly specifying parquet format."""
+    export_id = unique_export_id(request.node.name)
+
+    resp = client.export.create(
+        export_id=export_id,
+        backend=BACKEND,
+        file_format=ExportFileFormat.PARQUET,
+        include_collections=[COLLECTION_NAME],
+        wait_for_completion=True,
+    )
+    assert resp.status == ExportStatus.SUCCESS
+
+
+@pytest.mark.parametrize("include", [[COLLECTION_NAME], COLLECTION_NAME])
+def test_create_export_include_as_str_and_list(
+    client: weaviate.WeaviateClient, include: Union[str, List[str]], request: SubRequest
+) -> None:
+    """Verify include_collections accepts both str and list."""
+    export_id = unique_export_id(request.node.name)
+
+    resp = client.export.create(
+        export_id=export_id,
+        backend=BACKEND,
+        include_collections=include,
+        wait_for_completion=True,
+    )
+    assert resp.status == ExportStatus.SUCCESS
+    assert COLLECTION_NAME in resp.collections
+
+
+def test_cancel_export(client: weaviate.WeaviateClient, request: SubRequest) -> None:
+    """Cancel a running export."""
+    export_id = unique_export_id(request.node.name)
+
+    resp = client.export.create(
+        export_id=export_id,
+        backend=BACKEND,
+        include_collections=[COLLECTION_NAME],
+    )
+    assert resp.status in [ExportStatus.STARTED, ExportStatus.TRANSFERRING, ExportStatus.SUCCESS]
+
+    result = client.export.cancel(export_id=export_id, backend=BACKEND)
+    assert result is True
+
+    # verify it's cancelled or already completed (race condition)
+    start = time.time()
+    while time.time() - start < 5:
+        status = client.export.get_status(export_id=export_id, backend=BACKEND)
+        if status.status in [ExportStatus.CANCELLED, ExportStatus.SUCCESS]:
+            break
+        time.sleep(0.1)
+    assert status.status in [ExportStatus.CANCELLED, ExportStatus.SUCCESS]
+
+
+def test_fail_on_non_existing_collection(
+    client: weaviate.WeaviateClient, request: SubRequest
+) -> None:
+    """Fail export on non-existing collection."""
+    export_id = unique_export_id(request.node.name)
+    with pytest.raises(UnexpectedStatusCodeException):
+        client.export.create(
+            export_id=export_id,
+            backend=BACKEND,
+            include_collections=["NonExistingCollection"],
+            wait_for_completion=True,
+        )
+
+
+def test_fail_on_both_include_and_exclude(
+    client: weaviate.WeaviateClient, request: SubRequest
+) -> None:
+    """Fail when both include and exclude collections are set."""
+    export_id = unique_export_id(request.node.name)
+    with pytest.raises(TypeError):
+        client.export.create(
+            export_id=export_id,
+            backend=BACKEND,
+            include_collections=COLLECTION_NAME,
+            exclude_collections="SomeOther",
+        )
+
+
+def test_fail_status_for_non_existing_export(
+    client: weaviate.WeaviateClient, request: SubRequest
+) -> None:
+    """Fail checking status for non-existing export."""
+    export_id = unique_export_id(request.node.name)
+    with pytest.raises(UnexpectedStatusCodeException):
+        client.export.get_status(export_id=export_id, backend=BACKEND)
diff --git a/weaviate/classes/__init__.py b/weaviate/classes/__init__.py
index 467a17f37..d495744ac 100644
--- a/weaviate/classes/__init__.py
+++ b/weaviate/classes/__init__.py
@@ -5,6 +5,7 @@
     batch,
     config,
     data,
+    export,
     generate,
     generics,
     init,
@@ -22,6 +23,7 @@
     "config",
     "ConsistencyLevel",
     "data",
+    "export",
     "generate",
     "generics",
     "init",
diff --git a/weaviate/classes/export.py b/weaviate/classes/export.py
new file mode 100644
index 000000000..07e87a813
--- /dev/null
+++ b/weaviate/classes/export.py
@@ -0,0 +1,11 @@
+from weaviate.export.export import (
+    ExportConfig,
+    ExportFileFormat,
+    ExportStorage,
+)
+
+__all__ = [
+    "ExportConfig",
+    "ExportFileFormat",
+    "ExportStorage",
+]
diff --git a/weaviate/client.py b/weaviate/client.py
index d7f9080f4..f22389403 100644
--- a/weaviate/client.py
+++ b/weaviate/client.py
@@ -20,6 +20,7 @@
 from .connect.v4 import ConnectionAsync, ConnectionSync
 from .debug import _Debug, _DebugAsync
 from .embedded import EmbeddedOptions
+from .export import _Export, _ExportAsync
 from .groups import _Groups, _GroupsAsync
 from .rbac import _Roles, _RolesAsync
 from .types import NUMBER
@@ -76,6 +77,7 @@ def __init__(
         )
         self.alias = _AliasAsync(self._connection)
         self.backup = _BackupAsync(self._connection)
+        self.export = _ExportAsync(self._connection)
         self.batch = _BatchClientWrapperAsync(self._connection)
         self.cluster = _ClusterAsync(self._connection)
         self.collections = _CollectionsAsync(self._connection)
@@ -152,6 +154,7 @@ def __init__(
             consistency_level=None,
         )
         self.backup = _Backup(self._connection)
+        self.export = _Export(self._connection)
         self.cluster = _Cluster(self._connection)
         self.collections = collections
         self.debug = _Debug(self._connection)
diff --git a/weaviate/exceptions.py b/weaviate/exceptions.py
index 2a5b429d5..2b3d9ed09 100644
--- a/weaviate/exceptions.py
+++ b/weaviate/exceptions.py
@@ -141,6 +141,14 @@ class BackupCanceledError(WeaviateBaseError):
     """Backup canceled Exception."""
 
 
+class ExportFailedError(WeaviateBaseError):
+    """Export Failed Exception."""
+
+
+class ExportCancelledError(WeaviateBaseError):
+    """Export Cancelled Exception."""
+
+
 class EmptyResponseError(WeaviateBaseError):
     """Occurs when an HTTP request unexpectedly returns an empty response."""
 
diff --git a/weaviate/export/__init__.py b/weaviate/export/__init__.py
new file mode 100644
index 000000000..91de2d448
--- /dev/null
+++ b/weaviate/export/__init__.py
@@ -0,0 +1,7 @@
+"""Module for collection export operations."""
+
+from .async_ import _ExportAsync
+from .executor import ExportStorage
+from .sync import _Export
+
+__all__ = ["ExportStorage", "_ExportAsync", "_Export"]
diff --git a/weaviate/export/async_.py b/weaviate/export/async_.py
new file mode 100644
index 000000000..8bd1e3c44
--- /dev/null
+++ b/weaviate/export/async_.py
@@ -0,0 +1,8 @@
+from weaviate.connect import executor
+from weaviate.connect.v4 import ConnectionAsync
+from weaviate.export.executor import _ExportExecutor
+
+
+@executor.wrap("async")
+class _ExportAsync(_ExportExecutor[ConnectionAsync]):
+    pass
diff --git a/weaviate/export/async_.pyi b/weaviate/export/async_.pyi
new file mode 100644
index 000000000..3872a7cb9
--- /dev/null
+++ b/weaviate/export/async_.pyi
@@ -0,0 +1,38 @@
+from typing import List, Optional, Union
+
+from weaviate.connect.v4 import ConnectionAsync
+from weaviate.export.export import (
+    ExportConfig,
+    ExportCreateReturn,
+    ExportFileFormat,
+    ExportStatusReturn,
+    ExportStorage,
+)
+
+from .executor import _ExportExecutor
+
+class _ExportAsync(_ExportExecutor[ConnectionAsync]):
+    async def create(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        file_format: ExportFileFormat = ExportFileFormat.PARQUET,
+        include_collections: Union[List[str], str, None] = None,
+        exclude_collections: Union[List[str], str, None] = None,
+        wait_for_completion: bool = False,
+        config: Optional[ExportConfig] = None,
+    ) -> ExportCreateReturn: ...
+    async def get_status(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        bucket: Optional[str] = None,
+        path: Optional[str] = None,
+    ) -> ExportStatusReturn: ...
+    async def cancel(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        bucket: Optional[str] = None,
+        path: Optional[str] = None,
+    ) -> bool: ...
diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
new file mode 100644
index 000000000..25ecf643f
--- /dev/null
+++ b/weaviate/export/executor.py
@@ -0,0 +1,331 @@
+"""Export class definition."""
+
+import asyncio
+import time
+from typing import Dict, Generic, List, Optional, Tuple, Union
+
+from httpx import Response
+
+from weaviate.connect import executor
+from weaviate.connect.v4 import (
+    Connection,
+    ConnectionAsync,
+    ConnectionType,
+    _ExpectedStatusCodes,
+)
+from weaviate.exceptions import (
+    EmptyResponseException,
+    ExportCancelledError,
+    ExportFailedError,
+)
+from weaviate.export.export import (
+    STORAGE_NAMES,
+    ExportConfig,
+    ExportCreateReturn,
+    ExportFileFormat,
+    ExportStatus,
+    ExportStatusReturn,
+    ExportStorage,
+)
+from weaviate.util import (
+    _capitalize_first_letter,
+    _decode_json_response_dict,
+)
+
+
+class _ExportExecutor(Generic[ConnectionType]):
+    def __init__(self, connection: Connection):
+        self._connection = connection
+
+    def create(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        file_format: ExportFileFormat = ExportFileFormat.PARQUET,
+        include_collections: Union[List[str], str, None] = None,
+        exclude_collections: Union[List[str], str, None] = None,
+        wait_for_completion: bool = False,
+        config: Optional[ExportConfig] = None,
+    ) -> executor.Result[ExportCreateReturn]:
+        """Create an export of all/per collection Weaviate objects.
+
+        Args:
+            export_id: The identifier name of the export.
+            backend: The backend storage where to create the export.
+            file_format: The file format of the export. By default PARQUET.
+            include_collections: The collection/list of collections to be included in the export. If not specified all
+                collections will be included. Either `include_collections` or `exclude_collections` can be set.
+            exclude_collections: The collection/list of collections to be excluded in the export.
+                Either `include_collections` or `exclude_collections` can be set.
+            wait_for_completion: Whether to wait until the export is done. By default False.
+            config: The configuration of the export (bucket, path). By default None.
+
+        Returns:
+            An `ExportCreateReturn` object that contains the export creation response.
+
+        Raises:
+            weaviate.exceptions.UnexpectedStatusCodeError: If weaviate reports a non-OK status.
+            TypeError: One of the arguments have a wrong type.
+        """
+        (
+            export_id,
+            backend,
+            include_collections,
+            exclude_collections,
+        ) = _get_and_validate_create_arguments(
+            export_id=export_id,
+            backend=backend,
+            include_classes=include_collections,
+            exclude_classes=exclude_collections,
+            wait_for_completion=wait_for_completion,
+        )
+
+        payload: dict = {
+            "id": export_id,
+            "file_format": file_format.value,
+            "include": include_collections,
+            "exclude": exclude_collections,
+        }
+
+        if config is not None:
+            config_dict: Dict[str, str] = {}
+            if config.bucket is not None:
+                config_dict["bucket"] = config.bucket
+            if config.path is not None:
+                config_dict["path"] = config.path
+            if config_dict:
+                payload["config"] = config_dict
+
+        path = f"/export/{backend.value}"
+
+        if isinstance(self._connection, ConnectionAsync):
+
+            async def _execute() -> ExportCreateReturn:
+                res = await executor.aresult(
+                    self._connection.post(
+                        path=path,
+                        weaviate_object=payload,
+                        error_msg="Export creation failed due to connection error.",
+                    )
+                )
+                create_status = _decode_json_response_dict(res, "Export creation")
+                assert create_status is not None
+                if wait_for_completion:
+                    while True:
+                        status = await executor.aresult(
+                            self.get_status(
+                                export_id=export_id,
+                                backend=backend,
+                                bucket=config.bucket if config else None,
+                                path=config.path if config else None,
+                            )
+                        )
+                        create_status["status"] = status.status
+                        if status.status == ExportStatus.SUCCESS:
+                            break
+                        if status.status == ExportStatus.FAILED:
+                            raise ExportFailedError(
+                                f"Export failed: {create_status} with error: {status.error}"
+                            )
+                        if status.status == ExportStatus.CANCELLED:
+                            raise ExportCancelledError(
+                                f"Export was cancelled: {create_status} with error: {status.error}"
+                            )
+                        await asyncio.sleep(1)
+                return ExportCreateReturn(**create_status)
+
+            return _execute()
+
+        res = executor.result(
+            self._connection.post(
+                path=path,
+                weaviate_object=payload,
+                error_msg="Export creation failed due to connection error.",
+            )
+        )
+        create_status = _decode_json_response_dict(res, "Export creation")
+        assert create_status is not None
+        if wait_for_completion:
+            while True:
+                status = executor.result(
+                    self.get_status(
+                        export_id=export_id,
+                        backend=backend,
+                        bucket=config.bucket if config else None,
+                        path=config.path if config else None,
+                    )
+                )
+                create_status["status"] = status.status
+                if status.status == ExportStatus.SUCCESS:
+                    break
+                if status.status == ExportStatus.FAILED:
+                    raise ExportFailedError(
+                        f"Export failed: {create_status} with error: {status.error}"
+                    )
+                if status.status == ExportStatus.CANCELLED:
+                    raise ExportCancelledError(
+                        f"Export was cancelled: {create_status} with error: {status.error}"
+                    )
+                time.sleep(1)
+        return ExportCreateReturn(**create_status)
+
+    def get_status(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        bucket: Optional[str] = None,
+        path: Optional[str] = None,
+    ) -> executor.Result[ExportStatusReturn]:
+        """Check the status of an export.
+
+        Args:
+            export_id: The identifier name of the export.
+            backend: The backend storage where the export was created.
+            bucket: The bucket of the export location. By default None.
+            path: The path of the export location. By default None.
+
+        Returns:
+            An `ExportStatusReturn` object that contains the export status response.
+        """
+        export_id, backend = _get_and_validate_get_status(
+            export_id=export_id,
+            backend=backend,
+        )
+
+        url_path = f"/export/{backend.value}/{export_id}"
+        params: Dict[str, str] = {}
+        if bucket is not None:
+            params["bucket"] = bucket
+        if path is not None:
+            params["path"] = path
+
+        def resp(res: Response) -> ExportStatusReturn:
+            typed_response = _decode_json_response_dict(res, "Export status check")
+            if typed_response is None:
+                raise EmptyResponseException()
+            return ExportStatusReturn(**typed_response)
+
+        return executor.execute(
+            response_callback=resp,
+            method=self._connection.get,
+            path=url_path,
+            params=params,
+            error_msg="Export status check failed due to connection error.",
+        )
+
+    def cancel(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        bucket: Optional[str] = None,
+        path: Optional[str] = None,
+    ) -> executor.Result[bool]:
+        """Cancel a running export.
+
+        Args:
+            export_id: The identifier name of the export.
+            backend: The backend storage where the export was created.
+            bucket: The bucket of the export location. By default None.
+            path: The path of the export location. By default None.
+
+        Returns:
+            A bool indicating if the cancellation was successful.
+        """
+        export_id, backend = _get_and_validate_get_status(
+            export_id=export_id,
+            backend=backend,
+        )
+        url_path = f"/export/{backend.value}/{export_id}"
+        params: Dict[str, str] = {}
+        if bucket is not None:
+            params["bucket"] = bucket
+        if path is not None:
+            params["path"] = path
+
+        def resp(res: Response) -> bool:
+            if res.status_code == 204:
+                return True
+            typed_response = _decode_json_response_dict(res, "Export cancel")
+            if typed_response is None:
+                raise EmptyResponseException()
+            return False
+
+        return executor.execute(
+            response_callback=resp,
+            method=self._connection.delete,
+            path=url_path,
+            params=params,
+            error_msg="Export cancel failed due to connection error.",
+            status_codes=_ExpectedStatusCodes(ok_in=[204, 404], error="cancel export"),
+        )
+
+
+def _get_and_validate_create_arguments(
+    export_id: str,
+    backend: Union[str, ExportStorage],
+    include_classes: Union[List[str], str, None],
+    exclude_classes: Union[List[str], str, None],
+    wait_for_completion: bool,
+) -> Tuple[str, ExportStorage, List[str], List[str]]:
+    if not isinstance(export_id, str):
+        raise TypeError(f"'export_id' must be of type str. Given type: {type(export_id)}.")
+    if isinstance(backend, str):
+        try:
+            backend = ExportStorage(backend.lower())
+        except KeyError:
+            raise ValueError(
+                f"'backend' must have one of these values: {STORAGE_NAMES}. Given value: {backend}."
+            )
+
+    if not isinstance(wait_for_completion, bool):
+        raise TypeError(
+            f"'wait_for_completion' must be of type bool. Given type: {type(wait_for_completion)}."
+        )
+
+    if include_classes is not None:
+        if isinstance(include_classes, str):
+            include_classes = [include_classes]
+        elif not isinstance(include_classes, list):
+            raise TypeError(
+                "'include_collections' must be of type str, list of str or None. "
+                f"Given type: {type(include_classes)}."
+            )
+    else:
+        include_classes = []
+
+    if exclude_classes is not None:
+        if isinstance(exclude_classes, str):
+            exclude_classes = [exclude_classes]
+        elif not isinstance(exclude_classes, list):
+            raise TypeError(
+                "'exclude_collections' must be of type str, list of str or None. "
+                f"Given type: {type(exclude_classes)}."
+            )
+    else:
+        exclude_classes = []
+
+    if include_classes and exclude_classes:
+        raise TypeError(
+            "Either 'include_collections' OR 'exclude_collections' can be set, not both."
+        )
+
+    include_classes = [_capitalize_first_letter(cls) for cls in include_classes]
+    exclude_classes = [_capitalize_first_letter(cls) for cls in exclude_classes]
+
+    return (export_id, backend, include_classes, exclude_classes)
+
+
+def _get_and_validate_get_status(
+    export_id: str, backend: Union[str, ExportStorage]
+) -> Tuple[str, ExportStorage]:
+    if not isinstance(export_id, str):
+        raise TypeError(f"'export_id' must be of type str. Given type: {type(export_id)}.")
+    if isinstance(backend, str):
+        try:
+            backend = ExportStorage(backend.lower())
+        except KeyError:
+            raise ValueError(
+                f"'backend' must have one of these values: {STORAGE_NAMES}. Given value: {backend}."
+            )
+
+    return (export_id, backend)
diff --git a/weaviate/export/export.py b/weaviate/export/export.py
new file mode 100644
index 000000000..913f6198c
--- /dev/null
+++ b/weaviate/export/export.py
@@ -0,0 +1,92 @@
+"""Export models and enums."""
+
+from datetime import datetime
+from enum import Enum
+from typing import Dict, List, Optional
+
+from pydantic import BaseModel, Field
+
+STORAGE_NAMES = {
+    "filesystem",
+    "s3",
+    "gcs",
+    "azure",
+}
+
+
+class ExportStorage(str, Enum):
+    """Which backend should be used to write the export to."""
+
+    FILESYSTEM = "filesystem"
+    S3 = "s3"
+    GCS = "gcs"
+    AZURE = "azure"
+
+
+class ExportFileFormat(str, Enum):
+    """Which file format should be used for the export."""
+
+    PARQUET = "parquet"
+
+
+class ExportStatus(str, Enum):
+    """The status of an export."""
+
+    STARTED = "STARTED"
+    TRANSFERRING = "TRANSFERRING"
+    SUCCESS = "SUCCESS"
+    FAILED = "FAILED"
+    CANCELLED = "CANCELLED"
+    SKIPPED = "SKIPPED"
+
+
+class ShardExportStatus(str, Enum):
+    """The status of an individual shard export."""
+
+    STARTED = "STARTED"
+    TRANSFERRING = "TRANSFERRING"
+    SUCCESS = "SUCCESS"
+    FAILED = "FAILED"
+    CANCELLED = "CANCELLED"
+    SKIPPED = "SKIPPED"
+
+
+class ExportConfig(BaseModel):
+    """Configuration for where to write the export."""
+
+    bucket: Optional[str] = None
+    path: Optional[str] = None
+
+
+class ShardProgress(BaseModel):
+    """Progress of a single shard export."""
+
+    status: ShardExportStatus
+    objects_exported: int = Field(alias="objectsExported", default=0)
+    error: Optional[str] = None
+    skip_reason: Optional[str] = Field(alias="skipReason", default=None)
+
+    model_config = {"populate_by_name": True}
+
+
+class ExportCreateReturn(BaseModel):
+    """Return type of the export creation method."""
+
+    export_id: str = Field(alias="id")
+    backend: str
+    path: str
+    status: ExportStatus
+    started_at: Optional[datetime] = Field(alias="startedAt", default=None)
+    collections: List[str] = Field(default_factory=list, alias="classes")
+
+    model_config = {"populate_by_name": True}
+
+
+class ExportStatusReturn(ExportCreateReturn):
+    """Return type of the export status method."""
+
+    shard_status: Optional[Dict[str, Dict[str, ShardProgress]]] = Field(
+        alias="shardStatus", default=None
+    )
+    error: Optional[str] = None
+    took_in_ms: Optional[int] = Field(alias="tookInMs", default=None)
diff --git a/weaviate/export/sync.py b/weaviate/export/sync.py
new file mode 100644
index 000000000..0510601f8
--- /dev/null
+++ b/weaviate/export/sync.py
@@ -0,0 +1,8 @@
+from weaviate.connect import executor
+from weaviate.connect.v4 import ConnectionSync
+from weaviate.export.executor import _ExportExecutor
+
+
+@executor.wrap("sync")
+class _Export(_ExportExecutor[ConnectionSync]):
+    pass
diff --git a/weaviate/export/sync.pyi b/weaviate/export/sync.pyi
new file mode 100644
index 000000000..615c79cfc
--- /dev/null
+++ b/weaviate/export/sync.pyi
@@ -0,0 +1,38 @@
+from typing import List, Optional, Union
+
+from weaviate.connect.v4 import ConnectionSync
+from weaviate.export.export import (
+    ExportConfig,
+    ExportCreateReturn,
+    ExportFileFormat,
+    ExportStatusReturn,
+    ExportStorage,
+)
+
+from .executor import _ExportExecutor
+
+class _Export(_ExportExecutor[ConnectionSync]):
+    def create(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        file_format: ExportFileFormat = ExportFileFormat.PARQUET,
+        include_collections: Union[List[str], str, None] = None,
+        exclude_collections: Union[List[str], str, None] = None,
+        wait_for_completion: bool = False,
+        config: Optional[ExportConfig] = None,
+    ) -> ExportCreateReturn: ...
+    def get_status(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        bucket: Optional[str] = None,
+        path: Optional[str] = None,
+    ) -> ExportStatusReturn: ...
+    def cancel(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        bucket: Optional[str] = None,
+        path: Optional[str] = None,
+    ) -> bool: ...
diff --git a/weaviate/outputs/__init__.py b/weaviate/outputs/__init__.py
index 62193fc35..cd6176d93 100644
--- a/weaviate/outputs/__init__.py
+++ b/weaviate/outputs/__init__.py
@@ -1,4 +1,16 @@
-from . import aggregate, backup, batch, cluster, config, data, query, replication, tenants, users
+from . import (
+    aggregate,
+    backup,
+    batch,
+    cluster,
+    config,
+    data,
+    export,
+    query,
+    replication,
+    tenants,
+    users,
+)
 
 __all__ = [
     "aggregate",
@@ -7,6 +19,7 @@
     "cluster",
     "config",
     "data",
+    "export",
     "query",
     "replication",
     "tenants",
diff --git a/weaviate/outputs/export.py b/weaviate/outputs/export.py
new file mode 100644
index 000000000..531b715c4
--- /dev/null
+++ b/weaviate/outputs/export.py
@@ -0,0 +1,15 @@
+from weaviate.export.export import (
+    ExportCreateReturn,
+    ExportStatus,
+    ExportStatusReturn,
+    ShardExportStatus,
+    ShardProgress,
+)
+
+__all__ = [
+    "ExportCreateReturn",
+    "ExportStatus",
+    "ExportStatusReturn",
+    "ShardExportStatus",
+    "ShardProgress",
+]

From 92c3d1f3090f12de6737fb6462d88103392148f2 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Mon, 9 Mar 2026 16:59:55 +0100
Subject: [PATCH 59/99] Small cleanup after review

---
 integration/test_export.py  | 15 ++++++++-------
 weaviate/export/async_.pyi  |  2 +-
 weaviate/export/executor.py |  4 ++--
 weaviate/export/export.py   |  3 ---
 weaviate/export/sync.pyi    |  2 +-
 5 files changed, 12 insertions(+), 14 deletions(-)

diff --git a/integration/test_export.py b/integration/test_export.py
index 791360e3a..314c6b8fc 100644
--- a/integration/test_export.py
+++ b/integration/test_export.py
@@ -6,7 +6,6 @@
 from _pytest.fixtures import SubRequest
 
 import weaviate
-from weaviate.auth import Auth
 from weaviate.collections.classes.config import DataType, Property
 from weaviate.exceptions import UnexpectedStatusCodeException
 from weaviate.export.export import (
@@ -17,9 +16,6 @@
 
 from .conftest import _sanitize_collection_name
 
-RBAC_PORTS = (8093, 50065)
-RBAC_AUTH_CREDS = Auth.api_key("admin-key")
-
 pytestmark = pytest.mark.xdist_group(name="export")
 
 BACKEND = ExportStorage.FILESYSTEM
@@ -39,9 +35,7 @@
 
 @pytest.fixture(scope="module")
 def client() -> Generator[weaviate.WeaviateClient, None, None]:
-    client = weaviate.connect_to_local(
-        port=RBAC_PORTS[0], grpc_port=RBAC_PORTS[1], auth_credentials=RBAC_AUTH_CREDS
-    )
+    client = weaviate.connect_to_local()
     client.collections.delete(COLLECTION_NAME)
 
     col = client.collections.create(
@@ -73,6 +67,7 @@ def test_create_export_with_waiting(client: weaviate.WeaviateClient, request: Su
     resp = client.export.create(
         export_id=export_id,
         backend=BACKEND,
+        file_format=ExportFileFormat.PARQUET,
         include_collections=[COLLECTION_NAME],
         wait_for_completion=True,
     )
@@ -89,6 +84,7 @@ def test_create_export_without_waiting(
     resp = client.export.create(
         export_id=export_id,
         backend=BACKEND,
+        file_format=ExportFileFormat.PARQUET,
         include_collections=[COLLECTION_NAME],
     )
     assert resp.status in [ExportStatus.STARTED, ExportStatus.TRANSFERRING, ExportStatus.SUCCESS]
@@ -115,6 +111,7 @@ def test_get_export_status(client: weaviate.WeaviateClient, request: SubRequest)
     client.export.create(
         export_id=export_id,
         backend=BACKEND,
+        file_format=ExportFileFormat.PARQUET,
         include_collections=[COLLECTION_NAME],
         wait_for_completion=True,
     )
@@ -151,6 +148,7 @@ def test_create_export_include_as_str_and_list(
     resp = client.export.create(
         export_id=export_id,
         backend=BACKEND,
+        file_format=ExportFileFormat.PARQUET,
         include_collections=include,
         wait_for_completion=True,
     )
@@ -165,6 +163,7 @@ def test_cancel_export(client: weaviate.WeaviateClient, request: SubRequest) ->
     resp = client.export.create(
         export_id=export_id,
         backend=BACKEND,
+        file_format=ExportFileFormat.PARQUET,
         include_collections=[COLLECTION_NAME],
     )
     assert resp.status in [ExportStatus.STARTED, ExportStatus.TRANSFERRING, ExportStatus.SUCCESS]
@@ -191,6 +190,7 @@ def test_fail_on_non_existing_collection(
         client.export.create(
             export_id=export_id,
             backend=BACKEND,
+            file_format=ExportFileFormat.PARQUET,
             include_collections=["NonExistingCollection"],
             wait_for_completion=True,
         )
@@ -205,6 +205,7 @@ def test_fail_on_both_include_and_exclude(
         client.export.create(
             export_id=export_id,
             backend=BACKEND,
+            file_format=ExportFileFormat.PARQUET,
             include_collections=COLLECTION_NAME,
             exclude_collections="SomeOther",
         )
diff --git a/weaviate/export/async_.pyi b/weaviate/export/async_.pyi
index 3872a7cb9..652c26bee 100644
--- a/weaviate/export/async_.pyi
+++ b/weaviate/export/async_.pyi
@@ -16,7 +16,7 @@ class _ExportAsync(_ExportExecutor[ConnectionAsync]):
         self,
         export_id: str,
         backend: ExportStorage,
-        file_format: ExportFileFormat = ExportFileFormat.PARQUET,
+        file_format: ExportFileFormat,
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: bool = False,
diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
index 25ecf643f..39d586f2f 100644
--- a/weaviate/export/executor.py
+++ b/weaviate/export/executor.py
@@ -41,7 +41,7 @@ def create(
         self,
         export_id: str,
         backend: ExportStorage,
-        file_format: ExportFileFormat = ExportFileFormat.PARQUET,
+        file_format: ExportFileFormat,
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: bool = False,
@@ -52,7 +52,7 @@ def create(
         Args:
             export_id: The identifier name of the export.
             backend: The backend storage where to create the export.
-            file_format: The file format of the export. By default PARQUET.
+            file_format: The file format of the export (e.g. ExportFileFormat.PARQUET).
             include_collections: The collection/list of collections to be included in the export. If not specified all
                 collections will be included. Either `include_collections` or `exclude_collections` can be set.
             exclude_collections: The collection/list of collections to be excluded in the export.
diff --git a/weaviate/export/export.py b/weaviate/export/export.py
index 913f6198c..30f603bb5 100644
--- a/weaviate/export/export.py
+++ b/weaviate/export/export.py
@@ -37,17 +37,14 @@ class ExportStatus(str, Enum):
     SUCCESS = "SUCCESS"
     FAILED = "FAILED"
     CANCELLED = "CANCELLED"
-    SKIPPED = "SKIPPED"
 
 
 class ShardExportStatus(str, Enum):
     """The status of an individual shard export."""
 
-    STARTED = "STARTED"
     TRANSFERRING = "TRANSFERRING"
     SUCCESS = "SUCCESS"
     FAILED = "FAILED"
-    CANCELLED = "CANCELLED"
     SKIPPED = "SKIPPED"
 
 
diff --git a/weaviate/export/sync.pyi b/weaviate/export/sync.pyi
index 615c79cfc..93ed5f4fa 100644
--- a/weaviate/export/sync.pyi
+++ b/weaviate/export/sync.pyi
@@ -16,7 +16,7 @@ class _Export(_ExportExecutor[ConnectionSync]):
         self,
         export_id: str,
         backend: ExportStorage,
-        file_format: ExportFileFormat = ExportFileFormat.PARQUET,
+        file_format: ExportFileFormat,
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: bool = False,

From 3dc9259d395b55dea431bd37e3342f07d7836665 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 10 Mar 2026 07:31:13 +0100
Subject: [PATCH 60/99] Rename ENum

---
 integration/test_export.py  |  4 ++--
 weaviate/exceptions.py      |  4 ++--
 weaviate/export/executor.py | 14 +++++++-------
 weaviate/export/export.py   |  2 +-
 4 files changed, 12 insertions(+), 12 deletions(-)

diff --git a/integration/test_export.py b/integration/test_export.py
index 314c6b8fc..7748740c4 100644
--- a/integration/test_export.py
+++ b/integration/test_export.py
@@ -175,10 +175,10 @@ def test_cancel_export(client: weaviate.WeaviateClient, request: SubRequest) ->
     start = time.time()
     while time.time() - start < 5:
         status = client.export.get_status(export_id=export_id, backend=BACKEND)
-        if status.status in [ExportStatus.CANCELLED, ExportStatus.SUCCESS]:
+        if status.status in [ExportStatus.CANCELED, ExportStatus.SUCCESS]:
             break
         time.sleep(0.1)
-    assert status.status in [ExportStatus.CANCELLED, ExportStatus.SUCCESS]
+    assert status.status in [ExportStatus.CANCELED, ExportStatus.SUCCESS]
 
 
 def test_fail_on_non_existing_collection(
diff --git a/weaviate/exceptions.py b/weaviate/exceptions.py
index 2b3d9ed09..ce0fe6f7e 100644
--- a/weaviate/exceptions.py
+++ b/weaviate/exceptions.py
@@ -145,8 +145,8 @@ class ExportFailedError(WeaviateBaseError):
     """Export Failed Exception."""
 
 
-class ExportCancelledError(WeaviateBaseError):
-    """Export Cancelled Exception."""
+class ExportCanceledError(WeaviateBaseError):
+    """Export Canceled Exception."""
 
 
 class EmptyResponseError(WeaviateBaseError):
diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
index 39d586f2f..0a54c741a 100644
--- a/weaviate/export/executor.py
+++ b/weaviate/export/executor.py
@@ -15,7 +15,7 @@
 )
 from weaviate.exceptions import (
     EmptyResponseException,
-    ExportCancelledError,
+    ExportCanceledError,
     ExportFailedError,
 )
 from weaviate.export.export import (
@@ -127,9 +127,9 @@ async def _execute() -> ExportCreateReturn:
                             raise ExportFailedError(
                                 f"Export failed: {create_status} with error: {status.error}"
                             )
-                        if status.status == ExportStatus.CANCELLED:
-                            raise ExportCancelledError(
-                                f"Export was cancelled: {create_status} with error: {status.error}"
+                        if status.status == ExportStatus.CANCELED:
+                            raise ExportCanceledError(
+                                f"Export was canceled: {create_status} with error: {status.error}"
                             )
                         await asyncio.sleep(1)
                 return ExportCreateReturn(**create_status)
@@ -162,9 +162,9 @@ async def _execute() -> ExportCreateReturn:
                     raise ExportFailedError(
                         f"Export failed: {create_status} with error: {status.error}"
                     )
-                if status.status == ExportStatus.CANCELLED:
-                    raise ExportCancelledError(
-                        f"Export was cancelled: {create_status} with error: {status.error}"
+                if status.status == ExportStatus.CANCELED:
+                    raise ExportCanceledError(
+                        f"Export was canceled: {create_status} with error: {status.error}"
                     )
                 time.sleep(1)
         return ExportCreateReturn(**create_status)
diff --git a/weaviate/export/export.py b/weaviate/export/export.py
index 30f603bb5..43ee08d99 100644
--- a/weaviate/export/export.py
+++ b/weaviate/export/export.py
@@ -36,7 +36,7 @@ class ExportStatus(str, Enum):
     TRANSFERRING = "TRANSFERRING"
     SUCCESS = "SUCCESS"
     FAILED = "FAILED"
-    CANCELLED = "CANCELLED"
+    CANCELED = "CANCELED"
 
 
 class ShardExportStatus(str, Enum):

From c36e540a0c3c20c372f3a12a570de6a73fcf856c Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 31 Mar 2026 14:28:12 -0700
Subject: [PATCH 61/99] adapt to latest version

---
 weaviate/export/async_.pyi  | 12 ++----------
 weaviate/export/executor.py | 14 +-------------
 weaviate/export/export.py   |  2 +-
 weaviate/export/sync.pyi    | 12 ++----------
 4 files changed, 6 insertions(+), 34 deletions(-)

diff --git a/weaviate/export/async_.pyi b/weaviate/export/async_.pyi
index 652c26bee..df4b7d7ab 100644
--- a/weaviate/export/async_.pyi
+++ b/weaviate/export/async_.pyi
@@ -23,16 +23,8 @@ class _ExportAsync(_ExportExecutor[ConnectionAsync]):
         config: Optional[ExportConfig] = None,
     ) -> ExportCreateReturn: ...
     async def get_status(
-        self,
-        export_id: str,
-        backend: ExportStorage,
-        bucket: Optional[str] = None,
-        path: Optional[str] = None,
+        self, export_id: str, backend: ExportStorage, path: Optional[str] = None
     ) -> ExportStatusReturn: ...
     async def cancel(
-        self,
-        export_id: str,
-        backend: ExportStorage,
-        bucket: Optional[str] = None,
-        path: Optional[str] = None,
+        self, export_id: str, backend: ExportStorage, path: Optional[str] = None
     ) -> bool: ...
diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
index 0a54c741a..db013da2c 100644
--- a/weaviate/export/executor.py
+++ b/weaviate/export/executor.py
@@ -58,7 +58,7 @@ def create(
             exclude_collections: The collection/list of collections to be excluded in the export.
                 Either `include_collections` or `exclude_collections` can be set.
             wait_for_completion: Whether to wait until the export is done. By default False.
-            config: The configuration of the export (bucket, path). By default None.
+            config: The configuration of the export (path). By default None.
 
         Returns:
             An `ExportCreateReturn` object that contains the export creation response.
@@ -89,8 +89,6 @@ def create(
 
         if config is not None:
             config_dict: Dict[str, str] = {}
-            if config.bucket is not None:
-                config_dict["bucket"] = config.bucket
             if config.path is not None:
                 config_dict["path"] = config.path
             if config_dict:
@@ -116,7 +114,6 @@ async def _execute() -> ExportCreateReturn:
                             self.get_status(
                                 export_id=export_id,
                                 backend=backend,
-                                bucket=config.bucket if config else None,
                                 path=config.path if config else None,
                             )
                         )
@@ -151,7 +148,6 @@ async def _execute() -> ExportCreateReturn:
                     self.get_status(
                         export_id=export_id,
                         backend=backend,
-                        bucket=config.bucket if config else None,
                         path=config.path if config else None,
                     )
                 )
@@ -173,7 +169,6 @@ def get_status(
         self,
         export_id: str,
         backend: ExportStorage,
-        bucket: Optional[str] = None,
         path: Optional[str] = None,
     ) -> executor.Result[ExportStatusReturn]:
         """Check the status of an export.
@@ -181,7 +176,6 @@ def get_status(
         Args:
             export_id: The identifier name of the export.
             backend: The backend storage where the export was created.
-            bucket: The bucket of the export location. By default None.
             path: The path of the export location. By default None.
 
         Returns:
@@ -194,8 +188,6 @@ def get_status(
 
         url_path = f"/export/{backend.value}/{export_id}"
         params: Dict[str, str] = {}
-        if bucket is not None:
-            params["bucket"] = bucket
         if path is not None:
             params["path"] = path
 
@@ -217,7 +209,6 @@ def cancel(
         self,
         export_id: str,
         backend: ExportStorage,
-        bucket: Optional[str] = None,
         path: Optional[str] = None,
     ) -> executor.Result[bool]:
         """Cancel a running export.
@@ -225,7 +216,6 @@ def cancel(
         Args:
             export_id: The identifier name of the export.
             backend: The backend storage where the export was created.
-            bucket: The bucket of the export location. By default None.
             path: The path of the export location. By default None.
 
         Returns:
@@ -237,8 +227,6 @@ def cancel(
         )
         url_path = f"/export/{backend.value}/{export_id}"
         params: Dict[str, str] = {}
-        if bucket is not None:
-            params["bucket"] = bucket
         if path is not None:
             params["path"] = path
 
diff --git a/weaviate/export/export.py b/weaviate/export/export.py
index 43ee08d99..e41704a0f 100644
--- a/weaviate/export/export.py
+++ b/weaviate/export/export.py
@@ -51,7 +51,6 @@ class ShardExportStatus(str, Enum):
 class ExportConfig(BaseModel):
     """Configuration for where to write the export."""
 
-    bucket: Optional[str] = None
     path: Optional[str] = None
 
 
@@ -82,6 +81,7 @@ class ExportCreateReturn(BaseModel):
 class ExportStatusReturn(ExportCreateReturn):
     """Return type of the export status method."""
 
+    completed_at: Optional[datetime] = Field(alias="completedAt", default=None)
     shard_status: Optional[Dict[str, Dict[str, ShardProgress]]] = Field(
         alias="shardStatus", default=None
     )
diff --git a/weaviate/export/sync.pyi b/weaviate/export/sync.pyi
index 93ed5f4fa..fa00adb2c 100644
--- a/weaviate/export/sync.pyi
+++ b/weaviate/export/sync.pyi
@@ -23,16 +23,8 @@ class _Export(_ExportExecutor[ConnectionSync]):
         config: Optional[ExportConfig] = None,
     ) -> ExportCreateReturn: ...
     def get_status(
-        self,
-        export_id: str,
-        backend: ExportStorage,
-        bucket: Optional[str] = None,
-        path: Optional[str] = None,
+        self, export_id: str, backend: ExportStorage, path: Optional[str] = None
     ) -> ExportStatusReturn: ...
     def cancel(
-        self,
-        export_id: str,
-        backend: ExportStorage,
-        bucket: Optional[str] = None,
-        path: Optional[str] = None,
+        self, export_id: str, backend: ExportStorage, path: Optional[str] = None
     ) -> bool: ...

From 54eea32d0571e3be41aa1551d22fdcf7751224a6 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 31 Mar 2026 15:46:21 -0700
Subject: [PATCH 62/99] Update UX

---
 weaviate/export/async_.pyi  | 21 +++++++--
 weaviate/export/executor.py | 87 ++++++++++++++++++++++---------------
 weaviate/export/sync.pyi    | 21 +++++++--
 3 files changed, 87 insertions(+), 42 deletions(-)

diff --git a/weaviate/export/async_.pyi b/weaviate/export/async_.pyi
index df4b7d7ab..8450b1c1d 100644
--- a/weaviate/export/async_.pyi
+++ b/weaviate/export/async_.pyi
@@ -1,4 +1,4 @@
-from typing import List, Optional, Union
+from typing import List, Literal, Optional, Union, overload
 
 from weaviate.connect.v4 import ConnectionAsync
 from weaviate.export.export import (
@@ -12,6 +12,7 @@ from weaviate.export.export import (
 from .executor import _ExportExecutor
 
 class _ExportAsync(_ExportExecutor[ConnectionAsync]):
+    @overload
     async def create(
         self,
         export_id: str,
@@ -19,12 +20,24 @@ class _ExportAsync(_ExportExecutor[ConnectionAsync]):
         file_format: ExportFileFormat,
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
-        wait_for_completion: bool = False,
+        *,
+        wait_for_completion: Literal[True],
+        config: Optional[ExportConfig] = None,
+    ) -> ExportStatusReturn: ...
+    @overload
+    async def create(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        file_format: ExportFileFormat,
+        include_collections: Union[List[str], str, None] = None,
+        exclude_collections: Union[List[str], str, None] = None,
+        wait_for_completion: Literal[False] = False,
         config: Optional[ExportConfig] = None,
     ) -> ExportCreateReturn: ...
     async def get_status(
-        self, export_id: str, backend: ExportStorage, path: Optional[str] = None
+        self, export_id: str, backend: ExportStorage, config: Optional[ExportConfig] = None
     ) -> ExportStatusReturn: ...
     async def cancel(
-        self, export_id: str, backend: ExportStorage, path: Optional[str] = None
+        self, export_id: str, backend: ExportStorage, config: Optional[ExportConfig] = None
     ) -> bool: ...
diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
index db013da2c..587400a1c 100644
--- a/weaviate/export/executor.py
+++ b/weaviate/export/executor.py
@@ -2,7 +2,7 @@
 
 import asyncio
 import time
-from typing import Dict, Generic, List, Optional, Tuple, Union
+from typing import Dict, Generic, List, Literal, Optional, Tuple, Union, overload
 
 from httpx import Response
 
@@ -37,6 +37,31 @@ class _ExportExecutor(Generic[ConnectionType]):
     def __init__(self, connection: Connection):
         self._connection = connection
 
+    @overload
+    def create(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        file_format: ExportFileFormat,
+        include_collections: Union[List[str], str, None] = None,
+        exclude_collections: Union[List[str], str, None] = None,
+        *,
+        wait_for_completion: Literal[True],
+        config: Optional[ExportConfig] = None,
+    ) -> executor.Result[ExportStatusReturn]: ...
+
+    @overload
+    def create(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        file_format: ExportFileFormat,
+        include_collections: Union[List[str], str, None] = None,
+        exclude_collections: Union[List[str], str, None] = None,
+        wait_for_completion: Literal[False] = False,
+        config: Optional[ExportConfig] = None,
+    ) -> executor.Result[ExportCreateReturn]: ...
+
     def create(
         self,
         export_id: str,
@@ -46,7 +71,7 @@ def create(
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: bool = False,
         config: Optional[ExportConfig] = None,
-    ) -> executor.Result[ExportCreateReturn]:
+    ) -> executor.Result[Union[ExportCreateReturn, ExportStatusReturn]]:
         """Create an export of all/per collection Weaviate objects.
 
         Args:
@@ -83,9 +108,11 @@ def create(
         payload: dict = {
             "id": export_id,
             "file_format": file_format.value,
-            "include": include_collections,
-            "exclude": exclude_collections,
         }
+        if include_collections:
+            payload["include"] = include_collections
+        if exclude_collections:
+            payload["exclude"] = exclude_collections
 
         if config is not None:
             config_dict: Dict[str, str] = {}
@@ -98,7 +125,7 @@ def create(
 
         if isinstance(self._connection, ConnectionAsync):
 
-            async def _execute() -> ExportCreateReturn:
+            async def _execute() -> Union[ExportCreateReturn, ExportStatusReturn]:
                 res = await executor.aresult(
                     self._connection.post(
                         path=path,
@@ -114,19 +141,16 @@ async def _execute() -> ExportCreateReturn:
                             self.get_status(
                                 export_id=export_id,
                                 backend=backend,
-                                path=config.path if config else None,
+                                config=config,
                             )
                         )
-                        create_status["status"] = status.status
                         if status.status == ExportStatus.SUCCESS:
-                            break
+                            return status
                         if status.status == ExportStatus.FAILED:
-                            raise ExportFailedError(
-                                f"Export failed: {create_status} with error: {status.error}"
-                            )
+                            raise ExportFailedError(f"Export failed with error: {status.error}")
                         if status.status == ExportStatus.CANCELED:
                             raise ExportCanceledError(
-                                f"Export was canceled: {create_status} with error: {status.error}"
+                                f"Export was canceled with error: {status.error}"
                             )
                         await asyncio.sleep(1)
                 return ExportCreateReturn(**create_status)
@@ -148,20 +172,15 @@ async def _execute() -> ExportCreateReturn:
                     self.get_status(
                         export_id=export_id,
                         backend=backend,
-                        path=config.path if config else None,
+                        config=config,
                     )
                 )
-                create_status["status"] = status.status
                 if status.status == ExportStatus.SUCCESS:
-                    break
+                    return status
                 if status.status == ExportStatus.FAILED:
-                    raise ExportFailedError(
-                        f"Export failed: {create_status} with error: {status.error}"
-                    )
+                    raise ExportFailedError(f"Export failed with error: {status.error}")
                 if status.status == ExportStatus.CANCELED:
-                    raise ExportCanceledError(
-                        f"Export was canceled: {create_status} with error: {status.error}"
-                    )
+                    raise ExportCanceledError(f"Export was canceled with error: {status.error}")
                 time.sleep(1)
         return ExportCreateReturn(**create_status)
 
@@ -169,14 +188,14 @@ def get_status(
         self,
         export_id: str,
         backend: ExportStorage,
-        path: Optional[str] = None,
+        config: Optional[ExportConfig] = None,
     ) -> executor.Result[ExportStatusReturn]:
         """Check the status of an export.
 
         Args:
             export_id: The identifier name of the export.
             backend: The backend storage where the export was created.
-            path: The path of the export location. By default None.
+            config: The configuration of the export (path). By default None.
 
         Returns:
             An `ExportStatusReturn` object that contains the export status response.
@@ -188,8 +207,8 @@ def get_status(
 
         url_path = f"/export/{backend.value}/{export_id}"
         params: Dict[str, str] = {}
-        if path is not None:
-            params["path"] = path
+        if config is not None and config.path is not None:
+            params["path"] = config.path
 
         def resp(res: Response) -> ExportStatusReturn:
             typed_response = _decode_json_response_dict(res, "Export status check")
@@ -209,17 +228,17 @@ def cancel(
         self,
         export_id: str,
         backend: ExportStorage,
-        path: Optional[str] = None,
+        config: Optional[ExportConfig] = None,
     ) -> executor.Result[bool]:
         """Cancel a running export.
 
         Args:
             export_id: The identifier name of the export.
             backend: The backend storage where the export was created.
-            path: The path of the export location. By default None.
+            config: The configuration of the export (path). By default None.
 
         Returns:
-            A bool indicating if the cancellation was successful.
+            True if the export was cancelled, False if the export had already finished.
         """
         export_id, backend = _get_and_validate_get_status(
             export_id=export_id,
@@ -227,15 +246,15 @@ def cancel(
         )
         url_path = f"/export/{backend.value}/{export_id}"
         params: Dict[str, str] = {}
-        if path is not None:
-            params["path"] = path
+        if config is not None and config.path is not None:
+            params["path"] = config.path
 
         def resp(res: Response) -> bool:
             if res.status_code == 204:
                 return True
-            typed_response = _decode_json_response_dict(res, "Export cancel")
-            if typed_response is None:
-                raise EmptyResponseException()
+            # 409 means export already finished — not an error, just already done
+            if res.status_code == 409:
+                return False
             return False
 
         return executor.execute(
@@ -244,7 +263,7 @@ def resp(res: Response) -> bool:
             path=url_path,
             params=params,
             error_msg="Export cancel failed due to connection error.",
-            status_codes=_ExpectedStatusCodes(ok_in=[204, 404], error="cancel export"),
+            status_codes=_ExpectedStatusCodes(ok_in=[204, 409], error="cancel export"),
         )
 
 
diff --git a/weaviate/export/sync.pyi b/weaviate/export/sync.pyi
index fa00adb2c..cd8325b39 100644
--- a/weaviate/export/sync.pyi
+++ b/weaviate/export/sync.pyi
@@ -1,4 +1,4 @@
-from typing import List, Optional, Union
+from typing import List, Literal, Optional, Union, overload
 
 from weaviate.connect.v4 import ConnectionSync
 from weaviate.export.export import (
@@ -12,6 +12,7 @@ from weaviate.export.export import (
 from .executor import _ExportExecutor
 
 class _Export(_ExportExecutor[ConnectionSync]):
+    @overload
     def create(
         self,
         export_id: str,
@@ -19,12 +20,24 @@ class _Export(_ExportExecutor[ConnectionSync]):
         file_format: ExportFileFormat,
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
-        wait_for_completion: bool = False,
+        *,
+        wait_for_completion: Literal[True],
+        config: Optional[ExportConfig] = None,
+    ) -> ExportStatusReturn: ...
+    @overload
+    def create(
+        self,
+        export_id: str,
+        backend: ExportStorage,
+        file_format: ExportFileFormat,
+        include_collections: Union[List[str], str, None] = None,
+        exclude_collections: Union[List[str], str, None] = None,
+        wait_for_completion: Literal[False] = False,
         config: Optional[ExportConfig] = None,
     ) -> ExportCreateReturn: ...
     def get_status(
-        self, export_id: str, backend: ExportStorage, path: Optional[str] = None
+        self, export_id: str, backend: ExportStorage, config: Optional[ExportConfig] = None
     ) -> ExportStatusReturn: ...
     def cancel(
-        self, export_id: str, backend: ExportStorage, path: Optional[str] = None
+        self, export_id: str, backend: ExportStorage, config: Optional[ExportConfig] = None
     ) -> bool: ...

From 42bfc5c6550c5b4959a1aec39ed64cc89431ba69 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 7 Apr 2026 15:21:27 -0700
Subject: [PATCH 63/99] Remove export path parameter

---
 weaviate/classes/export.py  |  2 --
 weaviate/export/async_.pyi  | 13 +++----------
 weaviate/export/executor.py | 28 +---------------------------
 weaviate/export/export.py   |  6 ------
 weaviate/export/sync.pyi    | 13 +++----------
 5 files changed, 7 insertions(+), 55 deletions(-)

diff --git a/weaviate/classes/export.py b/weaviate/classes/export.py
index 07e87a813..d14fb3f07 100644
--- a/weaviate/classes/export.py
+++ b/weaviate/classes/export.py
@@ -1,11 +1,9 @@
 from weaviate.export.export import (
-    ExportConfig,
     ExportFileFormat,
     ExportStorage,
 )
 
 __all__ = [
-    "ExportConfig",
     "ExportFileFormat",
     "ExportStorage",
 ]
diff --git a/weaviate/export/async_.pyi b/weaviate/export/async_.pyi
index 8450b1c1d..4987a0cc6 100644
--- a/weaviate/export/async_.pyi
+++ b/weaviate/export/async_.pyi
@@ -1,8 +1,7 @@
-from typing import List, Literal, Optional, Union, overload
+from typing import List, Literal, Union, overload
 
 from weaviate.connect.v4 import ConnectionAsync
 from weaviate.export.export import (
-    ExportConfig,
     ExportCreateReturn,
     ExportFileFormat,
     ExportStatusReturn,
@@ -22,7 +21,6 @@ class _ExportAsync(_ExportExecutor[ConnectionAsync]):
         exclude_collections: Union[List[str], str, None] = None,
         *,
         wait_for_completion: Literal[True],
-        config: Optional[ExportConfig] = None,
     ) -> ExportStatusReturn: ...
     @overload
     async def create(
@@ -33,11 +31,6 @@ class _ExportAsync(_ExportExecutor[ConnectionAsync]):
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: Literal[False] = False,
-        config: Optional[ExportConfig] = None,
     ) -> ExportCreateReturn: ...
-    async def get_status(
-        self, export_id: str, backend: ExportStorage, config: Optional[ExportConfig] = None
-    ) -> ExportStatusReturn: ...
-    async def cancel(
-        self, export_id: str, backend: ExportStorage, config: Optional[ExportConfig] = None
-    ) -> bool: ...
+    async def get_status(self, export_id: str, backend: ExportStorage) -> ExportStatusReturn: ...
+    async def cancel(self, export_id: str, backend: ExportStorage) -> bool: ...
diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
index 587400a1c..18f2f9f72 100644
--- a/weaviate/export/executor.py
+++ b/weaviate/export/executor.py
@@ -2,7 +2,7 @@
 
 import asyncio
 import time
-from typing import Dict, Generic, List, Literal, Optional, Tuple, Union, overload
+from typing import Generic, List, Literal, Tuple, Union, overload
 
 from httpx import Response
 
@@ -20,7 +20,6 @@
 )
 from weaviate.export.export import (
     STORAGE_NAMES,
-    ExportConfig,
     ExportCreateReturn,
     ExportFileFormat,
     ExportStatus,
@@ -47,7 +46,6 @@ def create(
         exclude_collections: Union[List[str], str, None] = None,
         *,
         wait_for_completion: Literal[True],
-        config: Optional[ExportConfig] = None,
     ) -> executor.Result[ExportStatusReturn]: ...
 
     @overload
@@ -59,7 +57,6 @@ def create(
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: Literal[False] = False,
-        config: Optional[ExportConfig] = None,
     ) -> executor.Result[ExportCreateReturn]: ...
 
     def create(
@@ -70,7 +67,6 @@ def create(
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: bool = False,
-        config: Optional[ExportConfig] = None,
     ) -> executor.Result[Union[ExportCreateReturn, ExportStatusReturn]]:
         """Create an export of all/per collection Weaviate objects.
 
@@ -83,7 +79,6 @@ def create(
             exclude_collections: The collection/list of collections to be excluded in the export.
                 Either `include_collections` or `exclude_collections` can be set.
             wait_for_completion: Whether to wait until the export is done. By default False.
-            config: The configuration of the export (path). By default None.
 
         Returns:
             An `ExportCreateReturn` object that contains the export creation response.
@@ -114,13 +109,6 @@ def create(
         if exclude_collections:
             payload["exclude"] = exclude_collections
 
-        if config is not None:
-            config_dict: Dict[str, str] = {}
-            if config.path is not None:
-                config_dict["path"] = config.path
-            if config_dict:
-                payload["config"] = config_dict
-
         path = f"/export/{backend.value}"
 
         if isinstance(self._connection, ConnectionAsync):
@@ -141,7 +129,6 @@ async def _execute() -> Union[ExportCreateReturn, ExportStatusReturn]:
                             self.get_status(
                                 export_id=export_id,
                                 backend=backend,
-                                config=config,
                             )
                         )
                         if status.status == ExportStatus.SUCCESS:
@@ -172,7 +159,6 @@ async def _execute() -> Union[ExportCreateReturn, ExportStatusReturn]:
                     self.get_status(
                         export_id=export_id,
                         backend=backend,
-                        config=config,
                     )
                 )
                 if status.status == ExportStatus.SUCCESS:
@@ -188,14 +174,12 @@ def get_status(
         self,
         export_id: str,
         backend: ExportStorage,
-        config: Optional[ExportConfig] = None,
     ) -> executor.Result[ExportStatusReturn]:
         """Check the status of an export.
 
         Args:
             export_id: The identifier name of the export.
             backend: The backend storage where the export was created.
-            config: The configuration of the export (path). By default None.
 
         Returns:
             An `ExportStatusReturn` object that contains the export status response.
@@ -206,9 +190,6 @@ def get_status(
         )
 
         url_path = f"/export/{backend.value}/{export_id}"
-        params: Dict[str, str] = {}
-        if config is not None and config.path is not None:
-            params["path"] = config.path
 
         def resp(res: Response) -> ExportStatusReturn:
             typed_response = _decode_json_response_dict(res, "Export status check")
@@ -220,7 +201,6 @@ def resp(res: Response) -> ExportStatusReturn:
             response_callback=resp,
             method=self._connection.get,
             path=url_path,
-            params=params,
             error_msg="Export status check failed due to connection error.",
         )
 
@@ -228,14 +208,12 @@ def cancel(
         self,
         export_id: str,
         backend: ExportStorage,
-        config: Optional[ExportConfig] = None,
     ) -> executor.Result[bool]:
         """Cancel a running export.
 
         Args:
             export_id: The identifier name of the export.
             backend: The backend storage where the export was created.
-            config: The configuration of the export (path). By default None.
 
         Returns:
             True if the export was cancelled, False if the export had already finished.
@@ -245,9 +223,6 @@ def cancel(
             backend=backend,
         )
         url_path = f"/export/{backend.value}/{export_id}"
-        params: Dict[str, str] = {}
-        if config is not None and config.path is not None:
-            params["path"] = config.path
 
         def resp(res: Response) -> bool:
             if res.status_code == 204:
@@ -261,7 +236,6 @@ def resp(res: Response) -> bool:
             response_callback=resp,
             method=self._connection.delete,
             path=url_path,
-            params=params,
             error_msg="Export cancel failed due to connection error.",
             status_codes=_ExpectedStatusCodes(ok_in=[204, 409], error="cancel export"),
         )
diff --git a/weaviate/export/export.py b/weaviate/export/export.py
index e41704a0f..3cdc7e945 100644
--- a/weaviate/export/export.py
+++ b/weaviate/export/export.py
@@ -48,12 +48,6 @@ class ShardExportStatus(str, Enum):
     SKIPPED = "SKIPPED"
 
 
-class ExportConfig(BaseModel):
-    """Configuration for where to write the export."""
-
-    path: Optional[str] = None
-
-
 class ShardProgress(BaseModel):
     """Progress of a single shard export."""
 
diff --git a/weaviate/export/sync.pyi b/weaviate/export/sync.pyi
index cd8325b39..8b1820581 100644
--- a/weaviate/export/sync.pyi
+++ b/weaviate/export/sync.pyi
@@ -1,8 +1,7 @@
-from typing import List, Literal, Optional, Union, overload
+from typing import List, Literal, Union, overload
 
 from weaviate.connect.v4 import ConnectionSync
 from weaviate.export.export import (
-    ExportConfig,
     ExportCreateReturn,
     ExportFileFormat,
     ExportStatusReturn,
@@ -22,7 +21,6 @@ class _Export(_ExportExecutor[ConnectionSync]):
         exclude_collections: Union[List[str], str, None] = None,
         *,
         wait_for_completion: Literal[True],
-        config: Optional[ExportConfig] = None,
     ) -> ExportStatusReturn: ...
     @overload
     def create(
@@ -33,11 +31,6 @@ class _Export(_ExportExecutor[ConnectionSync]):
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: Literal[False] = False,
-        config: Optional[ExportConfig] = None,
     ) -> ExportCreateReturn: ...
-    def get_status(
-        self, export_id: str, backend: ExportStorage, config: Optional[ExportConfig] = None
-    ) -> ExportStatusReturn: ...
-    def cancel(
-        self, export_id: str, backend: ExportStorage, config: Optional[ExportConfig] = None
-    ) -> bool: ...
+    def get_status(self, export_id: str, backend: ExportStorage) -> ExportStatusReturn: ...
+    def cancel(self, export_id: str, backend: ExportStorage) -> bool: ...

From 2c749671e989079eb2bf2b626bf007925393fb9c Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Mon, 13 Apr 2026 06:38:39 +0200
Subject: [PATCH 64/99] Self-review of changes

---
 integration/test_export.py  |  2 +-
 weaviate/export/executor.py | 23 +++++++++++++++++++++--
 weaviate/export/export.py   | 16 ++--------------
 weaviate/export/sync.pyi    |  4 +++-
 weaviate/outputs/export.py  |  2 ++
 5 files changed, 29 insertions(+), 18 deletions(-)

diff --git a/integration/test_export.py b/integration/test_export.py
index 7748740c4..b393d153e 100644
--- a/integration/test_export.py
+++ b/integration/test_export.py
@@ -201,7 +201,7 @@ def test_fail_on_both_include_and_exclude(
 ) -> None:
     """Fail when both include and exclude collections are set."""
     export_id = unique_export_id(request.node.name)
-    with pytest.raises(TypeError):
+    with pytest.raises(ValueError):
         client.export.create(
             export_id=export_id,
             backend=BACKEND,
diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
index 18f2f9f72..84d2651f6 100644
--- a/weaviate/export/executor.py
+++ b/weaviate/export/executor.py
@@ -6,6 +6,7 @@
 
 from httpx import Response
 
+from weaviate.backup.backup import STORAGE_NAMES
 from weaviate.connect import executor
 from weaviate.connect.v4 import (
     Connection,
@@ -17,9 +18,9 @@
     EmptyResponseException,
     ExportCanceledError,
     ExportFailedError,
+    WeaviateUnsupportedFeatureError,
 )
 from weaviate.export.export import (
-    STORAGE_NAMES,
     ExportCreateReturn,
     ExportFileFormat,
     ExportStatus,
@@ -87,6 +88,12 @@ def create(
             weaviate.exceptions.UnexpectedStatusCodeError: If weaviate reports a non-OK status.
             TypeError: One of the arguments have a wrong type.
         """
+        if self._connection._weaviate_version.is_lower_than(1, 37, 0):
+            raise WeaviateUnsupportedFeatureError(
+                "Collection export",
+                str(self._connection._weaviate_version),
+                "1.37.0",
+            )
         (
             export_id,
             backend,
@@ -184,6 +191,12 @@ def get_status(
         Returns:
             An `ExportStatusReturn` object that contains the export status response.
         """
+        if self._connection._weaviate_version.is_lower_than(1, 37, 0):
+            raise WeaviateUnsupportedFeatureError(
+                "Collection export",
+                str(self._connection._weaviate_version),
+                "1.37.0",
+            )
         export_id, backend = _get_and_validate_get_status(
             export_id=export_id,
             backend=backend,
@@ -218,6 +231,12 @@ def cancel(
         Returns:
             True if the export was cancelled, False if the export had already finished.
         """
+        if self._connection._weaviate_version.is_lower_than(1, 37, 0):
+            raise WeaviateUnsupportedFeatureError(
+                "Collection export",
+                str(self._connection._weaviate_version),
+                "1.37.0",
+            )
         export_id, backend = _get_and_validate_get_status(
             export_id=export_id,
             backend=backend,
@@ -286,7 +305,7 @@ def _get_and_validate_create_arguments(
         exclude_classes = []
 
     if include_classes and exclude_classes:
-        raise TypeError(
+        raise ValueError(
             "Either 'include_collections' OR 'exclude_collections' can be set, not both."
         )
 
diff --git a/weaviate/export/export.py b/weaviate/export/export.py
index 3cdc7e945..fe94afa5c 100644
--- a/weaviate/export/export.py
+++ b/weaviate/export/export.py
@@ -6,21 +6,9 @@
 
 from pydantic import BaseModel, Field
 
-STORAGE_NAMES = {
-    "filesystem",
-    "s3",
-    "gcs",
-    "azure",
-}
+from weaviate.backup.backup import BackupStorage
 
-
-class ExportStorage(str, Enum):
-    """Which backend should be used to write the export to."""
-
-    FILESYSTEM = "filesystem"
-    S3 = "s3"
-    GCS = "gcs"
-    AZURE = "azure"
+ExportStorage = BackupStorage
 
 
 class ExportFileFormat(str, Enum):
diff --git a/weaviate/export/sync.pyi b/weaviate/export/sync.pyi
index 8b1820581..c9caee127 100644
--- a/weaviate/export/sync.pyi
+++ b/weaviate/export/sync.pyi
@@ -1,6 +1,8 @@
 from typing import List, Literal, Union, overload
 
-from weaviate.connect.v4 import ConnectionSync
+from weaviate.connect.v4 import (
+    ConnectionSync,
+)
 from weaviate.export.export import (
     ExportCreateReturn,
     ExportFileFormat,
diff --git a/weaviate/outputs/export.py b/weaviate/outputs/export.py
index 531b715c4..de0554de4 100644
--- a/weaviate/outputs/export.py
+++ b/weaviate/outputs/export.py
@@ -2,6 +2,7 @@
     ExportCreateReturn,
     ExportStatus,
     ExportStatusReturn,
+    ExportStorage,
     ShardExportStatus,
     ShardProgress,
 )
@@ -10,6 +11,7 @@
     "ExportCreateReturn",
     "ExportStatus",
     "ExportStatusReturn",
+    "ExportStorage",
     "ShardExportStatus",
     "ShardProgress",
 ]

From ed9f28879a7a1454ca02122d73a96677749fd70f Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Mon, 13 Apr 2026 06:59:34 +0200
Subject: [PATCH 65/99] Review fixes

---
 integration/test_export.py  | 3 +--
 weaviate/client.pyi         | 3 +++
 weaviate/export/executor.py | 7 ++++---
 3 files changed, 8 insertions(+), 5 deletions(-)

diff --git a/integration/test_export.py b/integration/test_export.py
index b393d153e..b885e9588 100644
--- a/integration/test_export.py
+++ b/integration/test_export.py
@@ -168,8 +168,7 @@ def test_cancel_export(client: weaviate.WeaviateClient, request: SubRequest) ->
     )
     assert resp.status in [ExportStatus.STARTED, ExportStatus.TRANSFERRING, ExportStatus.SUCCESS]
 
-    result = client.export.cancel(export_id=export_id, backend=BACKEND)
-    assert result is True
+    client.export.cancel(export_id=export_id, backend=BACKEND)
 
     # verify it's cancelled or already completed (race condition)
     start = time.time()
diff --git a/weaviate/client.pyi b/weaviate/client.pyi
index 9b32af15f..0ac79415c 100644
--- a/weaviate/client.pyi
+++ b/weaviate/client.pyi
@@ -20,6 +20,7 @@ from .backup import _Backup, _BackupAsync
 from .cluster import _Cluster, _ClusterAsync
 from .collections.batch.client import _BatchClientWrapper, _BatchClientWrapperAsync
 from .debug import _Debug, _DebugAsync
+from .export import _Export, _ExportAsync
 from .rbac import _Roles, _RolesAsync
 from .types import NUMBER
 
@@ -29,6 +30,7 @@ class WeaviateAsyncClient(_WeaviateClientExecutor[ConnectionAsync]):
     _connection: ConnectionAsync
     alias: _AliasAsync
     backup: _BackupAsync
+    export: _ExportAsync
     batch: _BatchClientWrapperAsync
     collections: _CollectionsAsync
     cluster: _ClusterAsync
@@ -52,6 +54,7 @@ class WeaviateClient(_WeaviateClientExecutor[ConnectionSync]):
     _connection: ConnectionSync
     alias: _Alias
     backup: _Backup
+    export: _Export
     batch: _BatchClientWrapper
     collections: _Collections
     cluster: _Cluster
diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
index 84d2651f6..6d772f969 100644
--- a/weaviate/export/executor.py
+++ b/weaviate/export/executor.py
@@ -82,7 +82,8 @@ def create(
             wait_for_completion: Whether to wait until the export is done. By default False.
 
         Returns:
-            An `ExportCreateReturn` object that contains the export creation response.
+            An `ExportCreateReturn` when `wait_for_completion=False`, or an `ExportStatusReturn`
+            when `wait_for_completion=True` and the export completes successfully.
 
         Raises:
             weaviate.exceptions.UnexpectedStatusCodeError: If weaviate reports a non-OK status.
@@ -272,7 +273,7 @@ def _get_and_validate_create_arguments(
     if isinstance(backend, str):
         try:
             backend = ExportStorage(backend.lower())
-        except KeyError:
+        except ValueError:
             raise ValueError(
                 f"'backend' must have one of these values: {STORAGE_NAMES}. Given value: {backend}."
             )
@@ -323,7 +324,7 @@ def _get_and_validate_get_status(
     if isinstance(backend, str):
         try:
             backend = ExportStorage(backend.lower())
-        except KeyError:
+        except ValueError:
             raise ValueError(
                 f"'backend' must have one of these values: {STORAGE_NAMES}. Given value: {backend}."
             )

From e9192f85f742011fb1dab7019ec15e71c730da61 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Mon, 13 Apr 2026 07:43:59 +0200
Subject: [PATCH 66/99] Add version guard for export integration tests

---
 integration/test_export.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/integration/test_export.py b/integration/test_export.py
index b885e9588..b72166998 100644
--- a/integration/test_export.py
+++ b/integration/test_export.py
@@ -36,6 +36,9 @@
 @pytest.fixture(scope="module")
 def client() -> Generator[weaviate.WeaviateClient, None, None]:
     client = weaviate.connect_to_local()
+    if client._connection._weaviate_version.is_lower_than(1, 37, 0):
+        client.close()
+        pytest.skip("Collection export is not supported in versions lower than 1.37.0")
     client.collections.delete(COLLECTION_NAME)
 
     col = client.collections.create(

From 338195add4a518d02137c9f43ec4edd47f36f96d Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 14 Apr 2026 07:50:33 +0200
Subject: [PATCH 67/99] Update to latest image

---
 .github/workflows/main.yaml | 2 +-
 ci/docker-compose.yml       | 2 ++
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index d3dfd1fa4..a7d6cca4d 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -28,7 +28,7 @@ env:
   WEAVIATE_134: 1.34.19
   WEAVIATE_135: 1.35.16-efdedfa
   WEAVIATE_136: 1.36.9-d905e6c
-  WEAVIATE_137: 1.37.0-rc.1-578c4eb
+  WEAVIATE_137: 1.37.0-rc.1-bc3891e
 
 jobs:
   lint-and-format:
diff --git a/ci/docker-compose.yml b/ci/docker-compose.yml
index 20840fafa..ddc92d0ed 100644
--- a/ci/docker-compose.yml
+++ b/ci/docker-compose.yml
@@ -31,6 +31,8 @@ services:
       DISABLE_LAZY_LOAD_SHARDS: 'true'
       GRPC_MAX_MESSAGE_SIZE: 100000000  # 100mb
       OBJECTS_TTL_DELETE_SCHEDULE: "@every 12h" # for objectTTL tests to work
+      EXPORT_ENABLED: 'true'
+      EXPORT_DEFAULT_PATH: "/var/lib/weaviate/exports"
 
   contextionary:
     environment:

From 90840ce1c8c5e0b6b7a46129f72efd58c17d69a2 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 14 Apr 2026 08:04:40 +0200
Subject: [PATCH 68/99] Lowercase export ID

---
 weaviate/export/executor.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
index 6d772f969..52d073e0a 100644
--- a/weaviate/export/executor.py
+++ b/weaviate/export/executor.py
@@ -270,6 +270,7 @@ def _get_and_validate_create_arguments(
 ) -> Tuple[str, ExportStorage, List[str], List[str]]:
     if not isinstance(export_id, str):
         raise TypeError(f"'export_id' must be of type str. Given type: {type(export_id)}.")
+    export_id = export_id.lower()
     if isinstance(backend, str):
         try:
             backend = ExportStorage(backend.lower())
@@ -321,6 +322,7 @@ def _get_and_validate_get_status(
 ) -> Tuple[str, ExportStorage]:
     if not isinstance(export_id, str):
         raise TypeError(f"'export_id' must be of type str. Given type: {type(export_id)}.")
+    export_id = export_id.lower()
     if isinstance(backend, str):
         try:
             backend = ExportStorage(backend.lower())

From 584f8a68bb2e495e950334fff805b6da112173ce Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Wed, 15 Apr 2026 07:47:55 +0200
Subject: [PATCH 69/99] Enforce kwargs for export

---
 weaviate/collections/collections/executor.py | 2 +-
 weaviate/collections/config/executor.py      | 2 +-
 weaviate/export/async_.pyi                   | 7 ++++---
 weaviate/export/executor.py                  | 6 +++++-
 weaviate/export/sync.pyi                     | 7 ++++---
 5 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/weaviate/collections/collections/executor.py b/weaviate/collections/collections/executor.py
index 644a0c844..2a733356c 100644
--- a/weaviate/collections/collections/executor.py
+++ b/weaviate/collections/collections/executor.py
@@ -44,13 +44,13 @@
     _check_references_generic,
 )
 from weaviate.collections.collection import Collection, CollectionAsync
+from weaviate.collections.config.executor import _any_property_has_text_analyzer
 from weaviate.connect import executor
 from weaviate.connect.v4 import (
     ConnectionAsync,
     ConnectionType,
     _ExpectedStatusCodes,
 )
-from weaviate.collections.config.executor import _any_property_has_text_analyzer
 from weaviate.exceptions import WeaviateInvalidInputError, WeaviateUnsupportedFeatureError
 from weaviate.util import _capitalize_first_letter, _decode_json_response_dict
 from weaviate.validator import _validate_input, _ValidateArgument
diff --git a/weaviate/collections/config/executor.py b/weaviate/collections/config/executor.py
index bd302c3c4..103ab70ac 100644
--- a/weaviate/collections/config/executor.py
+++ b/weaviate/collections/config/executor.py
@@ -5,8 +5,8 @@
     Generic,
     List,
     Literal,
-    Sequence,
     Optional,
+    Sequence,
     Tuple,
     Union,
     cast,
diff --git a/weaviate/export/async_.pyi b/weaviate/export/async_.pyi
index 4987a0cc6..4f1b82ced 100644
--- a/weaviate/export/async_.pyi
+++ b/weaviate/export/async_.pyi
@@ -14,17 +14,18 @@ class _ExportAsync(_ExportExecutor[ConnectionAsync]):
     @overload
     async def create(
         self,
+        *,
         export_id: str,
         backend: ExportStorage,
         file_format: ExportFileFormat,
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
-        *,
         wait_for_completion: Literal[True],
     ) -> ExportStatusReturn: ...
     @overload
     async def create(
         self,
+        *,
         export_id: str,
         backend: ExportStorage,
         file_format: ExportFileFormat,
@@ -32,5 +33,5 @@ class _ExportAsync(_ExportExecutor[ConnectionAsync]):
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: Literal[False] = False,
     ) -> ExportCreateReturn: ...
-    async def get_status(self, export_id: str, backend: ExportStorage) -> ExportStatusReturn: ...
-    async def cancel(self, export_id: str, backend: ExportStorage) -> bool: ...
+    async def get_status(self, *, export_id: str, backend: ExportStorage) -> ExportStatusReturn: ...
+    async def cancel(self, *, export_id: str, backend: ExportStorage) -> bool: ...
diff --git a/weaviate/export/executor.py b/weaviate/export/executor.py
index 52d073e0a..516273c73 100644
--- a/weaviate/export/executor.py
+++ b/weaviate/export/executor.py
@@ -40,18 +40,19 @@ def __init__(self, connection: Connection):
     @overload
     def create(
         self,
+        *,
         export_id: str,
         backend: ExportStorage,
         file_format: ExportFileFormat,
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
-        *,
         wait_for_completion: Literal[True],
     ) -> executor.Result[ExportStatusReturn]: ...
 
     @overload
     def create(
         self,
+        *,
         export_id: str,
         backend: ExportStorage,
         file_format: ExportFileFormat,
@@ -62,6 +63,7 @@ def create(
 
     def create(
         self,
+        *,
         export_id: str,
         backend: ExportStorage,
         file_format: ExportFileFormat,
@@ -180,6 +182,7 @@ async def _execute() -> Union[ExportCreateReturn, ExportStatusReturn]:
 
     def get_status(
         self,
+        *,
         export_id: str,
         backend: ExportStorage,
     ) -> executor.Result[ExportStatusReturn]:
@@ -220,6 +223,7 @@ def resp(res: Response) -> ExportStatusReturn:
 
     def cancel(
         self,
+        *,
         export_id: str,
         backend: ExportStorage,
     ) -> executor.Result[bool]:
diff --git a/weaviate/export/sync.pyi b/weaviate/export/sync.pyi
index c9caee127..afe7284e2 100644
--- a/weaviate/export/sync.pyi
+++ b/weaviate/export/sync.pyi
@@ -16,17 +16,18 @@ class _Export(_ExportExecutor[ConnectionSync]):
     @overload
     def create(
         self,
+        *,
         export_id: str,
         backend: ExportStorage,
         file_format: ExportFileFormat,
         include_collections: Union[List[str], str, None] = None,
         exclude_collections: Union[List[str], str, None] = None,
-        *,
         wait_for_completion: Literal[True],
     ) -> ExportStatusReturn: ...
     @overload
     def create(
         self,
+        *,
         export_id: str,
         backend: ExportStorage,
         file_format: ExportFileFormat,
@@ -34,5 +35,5 @@ class _Export(_ExportExecutor[ConnectionSync]):
         exclude_collections: Union[List[str], str, None] = None,
         wait_for_completion: Literal[False] = False,
     ) -> ExportCreateReturn: ...
-    def get_status(self, export_id: str, backend: ExportStorage) -> ExportStatusReturn: ...
-    def cancel(self, export_id: str, backend: ExportStorage) -> bool: ...
+    def get_status(self, *, export_id: str, backend: ExportStorage) -> ExportStatusReturn: ...
+    def cancel(self, *, export_id: str, backend: ExportStorage) -> bool: ...

From 96ca193e7507b0e40f6b3865e4afbdbb7af2d7b0 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Wed, 15 Apr 2026 08:13:28 +0200
Subject: [PATCH 70/99] Fix tests

---
 integration/test_export.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/integration/test_export.py b/integration/test_export.py
index b72166998..71e3a146f 100644
--- a/integration/test_export.py
+++ b/integration/test_export.py
@@ -104,7 +104,7 @@ def test_create_export_without_waiting(
             break
         time.sleep(0.1)
 
-    assert status.export_id == export_id
+    assert status.export_id.lower() == export_id.lower()
 
 
 def test_get_export_status(client: weaviate.WeaviateClient, request: SubRequest) -> None:
@@ -121,7 +121,7 @@ def test_get_export_status(client: weaviate.WeaviateClient, request: SubRequest)
 
     status = client.export.get_status(export_id=export_id, backend=BACKEND)
     assert status.status == ExportStatus.SUCCESS
-    assert status.export_id == export_id
+    assert status.export_id.lower() == export_id.lower()
     assert status.backend == BACKEND.value
 
 

From 594e8ee7c8d38f995ed4ce24e942be07a98be51d Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Wed, 15 Apr 2026 08:38:09 +0200
Subject: [PATCH 71/99] Fix tests

---
 integration/test_collection_query_profile.py | 87 ++++++++++----------
 1 file changed, 44 insertions(+), 43 deletions(-)

diff --git a/integration/test_collection_query_profile.py b/integration/test_collection_query_profile.py
index 92687129d..3b57e27b9 100644
--- a/integration/test_collection_query_profile.py
+++ b/integration/test_collection_query_profile.py
@@ -1,12 +1,14 @@
 import re
+from typing import Any
 
 import pytest
 
-import weaviate
-from weaviate.collections.classes.config import Configure, DataType, Property
+from weaviate.collections import Collection
+from weaviate.collections.classes.config import DataType, Property
 from weaviate.collections.classes.data import DataObject
-from weaviate.collections.classes.grpc import MetadataQuery
+from weaviate.collections.classes.grpc import GroupBy, MetadataQuery
 from weaviate.collections.classes.internal import SearchProfileReturn
+from integration.conftest import CollectionFactory
 
 GO_DURATION_RE = re.compile(r"[\d.]+(ns|µs|ms|s|m|h)")
 
@@ -28,26 +30,12 @@ def assert_common_profile(profile: SearchProfileReturn) -> None:
         assert isinstance(value, str) and value != ""
 
 
-@pytest.fixture(scope="module")
-def client():
-    client = weaviate.connect_to_local()
-    yield client
-    client.close()
-
-
-@pytest.fixture(scope="module")
-def collection_with_data(client: weaviate.WeaviateClient):
-    if client._connection._weaviate_version.is_lower_than(1, 36, 9):
-        pytest.skip("Query profiling requires Weaviate >= 1.36.9")
-    name = "TestQueryProfile"
-    client.collections.delete(name)
-    collection = client.collections.create(
-        name=name,
-        vectorizer_config=Configure.Vectorizer.none(),
-        properties=[
-            Property(name="text", data_type=DataType.TEXT),
-        ],
+def _create_and_populate(collection_factory: CollectionFactory) -> Collection[Any, Any]:
+    collection = collection_factory(
+        properties=[Property(name="text", data_type=DataType.TEXT)],
     )
+    if collection._connection._weaviate_version.is_lower_than(1, 36, 9):
+        pytest.skip("Query profiling requires Weaviate >= 1.36.9")
     collection.data.insert_many(
         [
             DataObject(properties={"text": "hello world"}, vector=[1.0, 0.0, 0.0]),
@@ -55,13 +43,13 @@ def collection_with_data(client: weaviate.WeaviateClient):
             DataObject(properties={"text": "foo bar baz"}, vector=[0.0, 0.0, 1.0]),
         ]
     )
-    yield collection
-    client.collections.delete(name)
+    return collection
 
 
-def test_fetch_objects_with_query_profile(collection_with_data):
+def test_fetch_objects_with_query_profile(collection_factory: CollectionFactory) -> None:
     """Test that query profiling works with fetch_objects (object lookup)."""
-    result = collection_with_data.query.fetch_objects(
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.fetch_objects(
         return_metadata=MetadataQuery(query_profile=True),
     )
     assert len(result.objects) == 3
@@ -76,9 +64,10 @@ def test_fetch_objects_with_query_profile(collection_with_data):
     assert_common_profile(shard.searches["object"])
 
 
-def test_near_vector_with_query_profile(collection_with_data):
+def test_near_vector_with_query_profile(collection_factory: CollectionFactory) -> None:
     """Test that query profiling works with near_vector search."""
-    result = collection_with_data.query.near_vector(
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.near_vector(
         near_vector=[1.0, 0.0, 0.0],
         return_metadata=MetadataQuery(query_profile=True, distance=True),
         limit=2,
@@ -107,9 +96,10 @@ def test_near_vector_with_query_profile(collection_with_data):
     assert_go_duration(vector_profile.details["objects_took"], "objects_took")
 
 
-def test_bm25_with_query_profile(collection_with_data):
+def test_bm25_with_query_profile(collection_factory: CollectionFactory) -> None:
     """Test that query profiling works with BM25 keyword search."""
-    result = collection_with_data.query.bm25(
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.bm25(
         query="hello",
         return_metadata=MetadataQuery(query_profile=True, score=True),
     )
@@ -135,9 +125,10 @@ def test_bm25_with_query_profile(collection_with_data):
     assert int(keyword_profile.details["kwd_6_res_count"]) >= 0
 
 
-def test_hybrid_with_query_profile(collection_with_data):
+def test_hybrid_with_query_profile(collection_factory: CollectionFactory) -> None:
     """Test that query profiling works with hybrid search (both vector and keyword)."""
-    result = collection_with_data.query.hybrid(
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.hybrid(
         query="hello",
         vector=[1.0, 0.0, 0.0],
         return_metadata=MetadataQuery(query_profile=True),
@@ -157,11 +148,12 @@ def test_hybrid_with_query_profile(collection_with_data):
     assert "kwd_method" in shard.searches["keyword"].details
 
 
-def test_near_vector_group_by_with_query_profile(collection_with_data):
-    """Test that query profiling works with group_by (mirrors C# QueryProfiling_NearText_GroupBy_Returns_Profile)."""
-    from weaviate.collections.classes.grpc import GroupBy
-
-    result = collection_with_data.query.near_vector(
+def test_near_vector_group_by_with_query_profile(
+    collection_factory: CollectionFactory,
+) -> None:
+    """Test that query profiling works with group_by."""
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.near_vector(
         near_vector=[1.0, 0.0, 0.0],
         return_metadata=MetadataQuery(query_profile=True),
         group_by=GroupBy(prop="text", objects_per_group=1, number_of_groups=3),
@@ -174,17 +166,23 @@ def test_near_vector_group_by_with_query_profile(collection_with_data):
     assert_common_profile(shard.searches["vector"])
 
 
-def test_no_query_profile_when_not_requested(collection_with_data):
+def test_no_query_profile_when_not_requested(
+    collection_factory: CollectionFactory,
+) -> None:
     """Test that query_profile is None when not requested."""
-    result = collection_with_data.query.fetch_objects(
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.fetch_objects(
         return_metadata=MetadataQuery(distance=True),
     )
     assert result.query_profile is None
 
 
-def test_query_profile_with_metadata_list(collection_with_data):
+def test_query_profile_with_metadata_list(
+    collection_factory: CollectionFactory,
+) -> None:
     """Test that query profiling works when using list-style metadata."""
-    result = collection_with_data.query.near_vector(
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.near_vector(
         near_vector=[1.0, 0.0, 0.0],
         return_metadata=["query_profile", "distance"],
         limit=2,
@@ -197,9 +195,12 @@ def test_query_profile_with_metadata_list(collection_with_data):
     assert_common_profile(shard.searches["vector"])
 
 
-def test_query_profile_details_are_strings(collection_with_data):
+def test_query_profile_details_are_strings(
+    collection_factory: CollectionFactory,
+) -> None:
     """Test that all detail keys and values are non-empty strings."""
-    result = collection_with_data.query.near_vector(
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.near_vector(
         near_vector=[1.0, 0.0, 0.0],
         return_metadata=MetadataQuery(query_profile=True),
         limit=1,

From b83a9488ca83e93846ee1f0714e1a3b37d95ee4a Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Wed, 15 Apr 2026 09:17:27 +0200
Subject: [PATCH 72/99] Add negative assertions

---
 integration/test_collection_query_profile.py | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/integration/test_collection_query_profile.py b/integration/test_collection_query_profile.py
index 3b57e27b9..0c68c9e69 100644
--- a/integration/test_collection_query_profile.py
+++ b/integration/test_collection_query_profile.py
@@ -61,6 +61,8 @@ def test_fetch_objects_with_query_profile(collection_factory: CollectionFactory)
     assert shard.node != ""
 
     assert "object" in shard.searches
+    assert "vector" not in shard.searches
+    assert "keyword" not in shard.searches
     assert_common_profile(shard.searches["object"])
 
 
@@ -78,6 +80,8 @@ def test_near_vector_with_query_profile(collection_factory: CollectionFactory) -
 
     shard = result.query_profile.shards[0]
     assert "vector" in shard.searches
+    assert "keyword" not in shard.searches
+    assert "object" not in shard.searches
     vector_profile = shard.searches["vector"]
     assert_common_profile(vector_profile)
 
@@ -108,6 +112,8 @@ def test_bm25_with_query_profile(collection_factory: CollectionFactory) -> None:
 
     shard = result.query_profile.shards[0]
     assert "keyword" in shard.searches
+    assert "vector" not in shard.searches
+    assert "object" not in shard.searches
     keyword_profile = shard.searches["keyword"]
     assert_common_profile(keyword_profile)
 
@@ -140,6 +146,7 @@ def test_hybrid_with_query_profile(collection_factory: CollectionFactory) -> Non
     shard = result.query_profile.shards[0]
     assert "vector" in shard.searches, "Hybrid should produce a 'vector' profile"
     assert "keyword" in shard.searches, "Hybrid should produce a 'keyword' profile"
+    assert "object" not in shard.searches
 
     assert_common_profile(shard.searches["vector"])
     assert "vector_search_took" in shard.searches["vector"].details

From 8b2caaf7356223370b3eba7665d7c5e59c685be1 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Wed, 15 Apr 2026 10:39:17 +0100
Subject: [PATCH 73/99] refactor: names don't shadow existing

---
 integration/test_tokenize.py                  | 38 +++++++++----------
 weaviate/__init__.py                          |  4 +-
 weaviate/client.py                            |  6 +--
 weaviate/client.pyi                           |  6 +--
 weaviate/tokenization/__init__.py             |  7 ++++
 weaviate/tokenization/async_.py               |  8 ++++
 .../{tokenize => tokenization}/executor.py    |  6 +--
 .../types.py => tokenization/models.py}       |  0
 weaviate/tokenization/sync.py                 |  8 ++++
 weaviate/tokenize/__init__.py                 |  7 ----
 weaviate/tokenize/async_.py                   |  8 ----
 weaviate/tokenize/sync.py                     |  8 ----
 12 files changed, 53 insertions(+), 53 deletions(-)
 create mode 100644 weaviate/tokenization/__init__.py
 create mode 100644 weaviate/tokenization/async_.py
 rename weaviate/{tokenize => tokenization}/executor.py (97%)
 rename weaviate/{tokenize/types.py => tokenization/models.py} (100%)
 create mode 100644 weaviate/tokenization/sync.py
 delete mode 100644 weaviate/tokenize/__init__.py
 delete mode 100644 weaviate/tokenize/async_.py
 delete mode 100644 weaviate/tokenize/sync.py

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index e54f9d49d..b3ecff875 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -22,7 +22,7 @@
     _TextAnalyzerConfigCreate,
 )
 from weaviate.config import AdditionalConfig
-from weaviate.tokenize.types import TokenizeResult
+from weaviate.tokenization.models import TokenizeResult
 
 
 @pytest.fixture(scope="module")
@@ -69,19 +69,19 @@ def test_tokenization_enum(
         text: str,
         expected_tokens: list,
     ) -> None:
-        result = client.tokenize.text(text=text, tokenization=tokenization)
+        result = client.tokenization.text(text=text, tokenization=tokenization)
         assert isinstance(result, TokenizeResult)
         assert result.tokenization == tokenization.value
         assert result.indexed == expected_tokens
         assert result.query == expected_tokens
 
     def test_tokenization_string(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenize.text(text="hello world", tokenization="word")
+        result = client.tokenization.text(text="hello world", tokenization="word")
         assert result.tokenization == "word"
         assert result.indexed == ["hello", "world"]
 
     def test_stopword_preset_enum(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenize.text(
+        result = client.tokenization.text(
             text="The quick brown fox",
             tokenization=Tokenization.WORD,
             stopword_preset=StopwordsPreset.EN,
@@ -90,7 +90,7 @@ def test_stopword_preset_enum(self, client: weaviate.WeaviateClient) -> None:
         assert "quick" in result.query
 
     def test_stopword_preset_string(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenize.text(
+        result = client.tokenization.text(
             text="The quick brown fox",
             tokenization=Tokenization.WORD,
             stopword_preset="en",
@@ -98,7 +98,7 @@ def test_stopword_preset_string(self, client: weaviate.WeaviateClient) -> None:
         assert "the" not in result.query
 
     def test_ascii_fold_via_kwargs(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenize.text(
+        result = client.tokenization.text(
             text="L'école est fermée",
             tokenization=Tokenization.WORD,
             ascii_fold=True,
@@ -107,7 +107,7 @@ def test_ascii_fold_via_kwargs(self, client: weaviate.WeaviateClient) -> None:
 
     def test_ascii_fold_via_analyzer_config(self, client: weaviate.WeaviateClient) -> None:
         cfg = _TextAnalyzerConfigCreate(ascii_fold=True)
-        result = client.tokenize.text(
+        result = client.tokenization.text(
             text="L'école est fermée",
             tokenization=Tokenization.WORD,
             analyzer_config=cfg,
@@ -121,12 +121,12 @@ def test_analyzer_config_and_kwargs_produce_same_result(
         cfg = _TextAnalyzerConfigCreate(
             ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset=StopwordsPreset.EN
         )
-        via_config = client.tokenize.text(
+        via_config = client.tokenization.text(
             text="L'école est fermée",
             tokenization=Tokenization.WORD,
             analyzer_config=cfg,
         )
-        via_kwargs = client.tokenize.text(
+        via_kwargs = client.tokenization.text(
             text="L'école est fermée",
             tokenization=Tokenization.WORD,
             ascii_fold=True,
@@ -137,7 +137,7 @@ def test_analyzer_config_and_kwargs_produce_same_result(
         assert via_config.query == via_kwargs.query
 
     def test_stopword_presets_serialization(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenize.text(
+        result = client.tokenization.text(
             text="hello world test",
             tokenization=Tokenization.WORD,
             stopword_preset="custom",
@@ -149,7 +149,7 @@ def test_stopword_presets_serialization(self, client: weaviate.WeaviateClient) -
         assert result.query == ["hello", "world"]
 
     def test_stopword_presets_with_base_and_removals(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenize.text(
+        result = client.tokenization.text(
             text="the quick",
             tokenization=Tokenization.WORD,
             stopword_preset="en-no-the",
@@ -172,13 +172,13 @@ class TestDeserialization:
     """Verify the client correctly deserializes response fields into typed objects."""
 
     def test_result_type(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenize.text(text="hello", tokenization=Tokenization.WORD)
+        result = client.tokenization.text(text="hello", tokenization=Tokenization.WORD)
         assert isinstance(result, TokenizeResult)
         assert isinstance(result.indexed, list)
         assert isinstance(result.query, list)
 
     def test_analyzer_config_deserialized(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenize.text(
+        result = client.tokenization.text(
             text="L'école",
             tokenization=Tokenization.WORD,
             ascii_fold=True,
@@ -191,7 +191,7 @@ def test_analyzer_config_deserialized(self, client: weaviate.WeaviateClient) ->
         assert result.analyzer_config.stopword_preset == "en"
 
     def test_no_analyzer_config_returns_none(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenize.text(text="hello", tokenization=Tokenization.WORD)
+        result = client.tokenization.text(text="hello", tokenization=Tokenization.WORD)
         assert result.analyzer_config is None
 
     def test_stopword_config_deserialized_on_property(
@@ -214,7 +214,7 @@ def test_stopword_config_deserialized_on_property(
                     ],
                 }
             )
-            result = client.tokenize.property(
+            result = client.tokenization.for_property(
                 collection_name="TestDeserStopword",
                 property_name="title",
                 text="the quick",
@@ -243,7 +243,7 @@ def test_property_result_types(self, client: weaviate.WeaviateClient) -> None:
                     ],
                 }
             )
-            result = client.tokenize.property(
+            result = client.tokenization.for_property(
                 collection_name="TestDeserPropTypes",
                 property_name="tag",
                 text="  Hello World  ",
@@ -302,7 +302,7 @@ class TestAsyncClient:
 
     @pytest.mark.asyncio
     async def test_text_tokenize(self, async_client: weaviate.WeaviateAsyncClient) -> None:
-        result = await async_client.tokenize.text(
+        result = await async_client.tokenization.text(
             text="The quick brown fox",
             tokenization=Tokenization.WORD,
         )
@@ -314,7 +314,7 @@ async def test_text_with_analyzer_config(
         self, async_client: weaviate.WeaviateAsyncClient
     ) -> None:
         cfg = _TextAnalyzerConfigCreate(ascii_fold=True, stopword_preset=StopwordsPreset.EN)
-        result = await async_client.tokenize.text(
+        result = await async_client.tokenization.text(
             text="L'école est fermée",
             tokenization=Tokenization.WORD,
             analyzer_config=cfg,
@@ -341,7 +341,7 @@ async def test_property_tokenize(self, async_client: weaviate.WeaviateAsyncClien
                     ],
                 }
             )
-            result = await async_client.tokenize.property(
+            result = await async_client.tokenization.for_property(
                 collection_name="TestAsyncPropTokenize",
                 property_name="title",
                 text="The quick brown fox",
diff --git a/weaviate/__init__.py b/weaviate/__init__.py
index 6fd9368ea..2e7e5e58b 100644
--- a/weaviate/__init__.py
+++ b/weaviate/__init__.py
@@ -21,7 +21,7 @@
     embedded,
     exceptions,
     outputs,
-    tokenize,
+    tokenization,
     types,
 )
 from .client import Client, WeaviateAsyncClient, WeaviateClient
@@ -68,7 +68,7 @@
     "embedded",
     "exceptions",
     "outputs",
-    "tokenize",
+    "tokenization",
     "types",
     "use_async_with_custom",
     "use_async_with_embedded",
diff --git a/weaviate/client.py b/weaviate/client.py
index cbd12be9a..10ce4c77b 100644
--- a/weaviate/client.py
+++ b/weaviate/client.py
@@ -22,7 +22,7 @@
 from .embedded import EmbeddedOptions
 from .groups import _Groups, _GroupsAsync
 from .rbac import _Roles, _RolesAsync
-from .tokenize import _Tokenize, _TokenizeAsync
+from .tokenization import _Tokenization, _TokenizationAsync
 from .types import NUMBER
 from .users import _Users, _UsersAsync
 
@@ -83,7 +83,7 @@ def __init__(
         self.debug = _DebugAsync(self._connection)
         self.groups = _GroupsAsync(self._connection)
         self.roles = _RolesAsync(self._connection)
-        self.tokenize = _TokenizeAsync(self._connection)
+        self.tokenization = _TokenizationAsync(self._connection)
         self.users = _UsersAsync(self._connection)
 
     async def __aenter__(self) -> "WeaviateAsyncClient":
@@ -159,7 +159,7 @@ def __init__(
         self.debug = _Debug(self._connection)
         self.groups = _Groups(self._connection)
         self.roles = _Roles(self._connection)
-        self.tokenize = _Tokenize(self._connection)
+        self.tokenization = _Tokenization(self._connection)
         self.users = _Users(self._connection)
 
     def __enter__(self) -> "WeaviateClient":
diff --git a/weaviate/client.pyi b/weaviate/client.pyi
index a6a44f8f7..8fafdc3d1 100644
--- a/weaviate/client.pyi
+++ b/weaviate/client.pyi
@@ -21,7 +21,7 @@ from .cluster import _Cluster, _ClusterAsync
 from .collections.batch.client import _BatchClientWrapper, _BatchClientWrapperAsync
 from .debug import _Debug, _DebugAsync
 from .rbac import _Roles, _RolesAsync
-from .tokenize import _Tokenize, _TokenizeAsync
+from .tokenization import _Tokenization, _TokenizationAsync
 from .types import NUMBER
 
 TIMEOUT_TYPE = Union[Tuple[NUMBER, NUMBER], NUMBER]
@@ -36,7 +36,7 @@ class WeaviateAsyncClient(_WeaviateClientExecutor[ConnectionAsync]):
     debug: _DebugAsync
     groups: _GroupsAsync
     roles: _RolesAsync
-    tokenize: _TokenizeAsync
+    tokenization: _TokenizationAsync
     users: _UsersAsync
 
     async def close(self) -> None: ...
@@ -60,7 +60,7 @@ class WeaviateClient(_WeaviateClientExecutor[ConnectionSync]):
     debug: _Debug
     groups: _Groups
     roles: _Roles
-    tokenize: _Tokenize
+    tokenization: _Tokenization
     users: _Users
 
     def close(self) -> None: ...
diff --git a/weaviate/tokenization/__init__.py b/weaviate/tokenization/__init__.py
new file mode 100644
index 000000000..2437f7745
--- /dev/null
+++ b/weaviate/tokenization/__init__.py
@@ -0,0 +1,7 @@
+"""Module for tokenization operations."""
+
+from .async_ import _TokenizationAsync
+from .sync import _Tokenization
+from .models import TokenizeResult
+
+__all__ = ["_Tokenization", "_TokenizationAsync", "TokenizeResult"]
diff --git a/weaviate/tokenization/async_.py b/weaviate/tokenization/async_.py
new file mode 100644
index 000000000..5406a39dd
--- /dev/null
+++ b/weaviate/tokenization/async_.py
@@ -0,0 +1,8 @@
+from weaviate.connect import executor
+from weaviate.connect.v4 import ConnectionAsync
+from weaviate.tokenization.executor import _TokenizationExecutor
+
+
+@executor.wrap("async")
+class _TokenizationAsync(_TokenizationExecutor[ConnectionAsync]):
+    pass
diff --git a/weaviate/tokenize/executor.py b/weaviate/tokenization/executor.py
similarity index 97%
rename from weaviate/tokenize/executor.py
rename to weaviate/tokenization/executor.py
index bd2c24dc1..9ddf5f7ed 100644
--- a/weaviate/tokenize/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -14,7 +14,7 @@
 )
 from weaviate.connect import executor
 from weaviate.connect.v4 import ConnectionType, _ExpectedStatusCodes
-from weaviate.tokenize.types import TokenizeResult
+from weaviate.tokenization.models import TokenizeResult
 
 
 def _parse_analyzer_config(body: Dict[str, Any]) -> Optional[TextAnalyzerConfig]:
@@ -51,7 +51,7 @@ def _parse_tokenize_result(body: Dict[str, Any]) -> TokenizeResult:
     )
 
 
-class _TokenizeExecutor(Generic[ConnectionType]):
+class _TokenizationExecutor(Generic[ConnectionType]):
     def __init__(self, connection: ConnectionType):
         self._connection = connection
 
@@ -133,7 +133,7 @@ def resp(response: Response) -> TokenizeResult:
             status_codes=_ExpectedStatusCodes(ok_in=[200], error="tokenize text"),
         )
 
-    def property(
+    def for_property(
         self,
         collection_name: str,
         property_name: str,
diff --git a/weaviate/tokenize/types.py b/weaviate/tokenization/models.py
similarity index 100%
rename from weaviate/tokenize/types.py
rename to weaviate/tokenization/models.py
diff --git a/weaviate/tokenization/sync.py b/weaviate/tokenization/sync.py
new file mode 100644
index 000000000..ab28cc98e
--- /dev/null
+++ b/weaviate/tokenization/sync.py
@@ -0,0 +1,8 @@
+from weaviate.connect import executor
+from weaviate.connect.v4 import ConnectionSync
+from weaviate.tokenization.executor import _TokenizationExecutor
+
+
+@executor.wrap("sync")
+class _Tokenization(_TokenizationExecutor[ConnectionSync]):
+    pass
diff --git a/weaviate/tokenize/__init__.py b/weaviate/tokenize/__init__.py
deleted file mode 100644
index d0c2883c5..000000000
--- a/weaviate/tokenize/__init__.py
+++ /dev/null
@@ -1,7 +0,0 @@
-"""Module for tokenize operations."""
-
-from .async_ import _TokenizeAsync
-from .sync import _Tokenize
-from .types import TokenizeResult
-
-__all__ = ["_Tokenize", "_TokenizeAsync", "TokenizeResult"]
diff --git a/weaviate/tokenize/async_.py b/weaviate/tokenize/async_.py
deleted file mode 100644
index a59c392ea..000000000
--- a/weaviate/tokenize/async_.py
+++ /dev/null
@@ -1,8 +0,0 @@
-from weaviate.connect import executor
-from weaviate.connect.v4 import ConnectionAsync
-from weaviate.tokenize.executor import _TokenizeExecutor
-
-
-@executor.wrap("async")
-class _TokenizeAsync(_TokenizeExecutor[ConnectionAsync]):
-    pass
diff --git a/weaviate/tokenize/sync.py b/weaviate/tokenize/sync.py
deleted file mode 100644
index 755c42559..000000000
--- a/weaviate/tokenize/sync.py
+++ /dev/null
@@ -1,8 +0,0 @@
-from weaviate.connect import executor
-from weaviate.connect.v4 import ConnectionSync
-from weaviate.tokenize.executor import _TokenizeExecutor
-
-
-@executor.wrap("sync")
-class _Tokenize(_TokenizeExecutor[ConnectionSync]):
-    pass

From ede0b96477aa44f5de9fb964e8a48a84e126b408 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Wed, 15 Apr 2026 10:47:01 +0100
Subject: [PATCH 74/99] fix: add version gate

---
 integration/test_tokenize.py      | 24 ++++++++++++++++++++++++
 weaviate/tokenization/executor.py | 17 +++++++++++++++++
 2 files changed, 41 insertions(+)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index b3ecff875..47321aaf5 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -22,6 +22,7 @@
     _TextAnalyzerConfigCreate,
 )
 from weaviate.config import AdditionalConfig
+from weaviate.exceptions import WeaviateUnsupportedFeatureError
 from weaviate.tokenization.models import TokenizeResult
 
 
@@ -292,6 +293,29 @@ def test_empty_config_is_valid(self) -> None:
         assert cfg.stopwordPreset is None
 
 
+# ---------------------------------------------------------------------------
+# Version gate
+# ---------------------------------------------------------------------------
+
+
+class TestVersionGate:
+    """On Weaviate < 1.37 the client must raise before sending the request."""
+
+    def test_text_raises_on_old_server(self, client: weaviate.WeaviateClient) -> None:
+        if client._connection._weaviate_version.is_at_least(1, 37, 0):
+            pytest.skip("Version gate only applies to Weaviate < 1.37.0")
+        with pytest.raises(WeaviateUnsupportedFeatureError):
+            client.tokenization.text(text="hello", tokenization=Tokenization.WORD)
+
+    def test_for_property_raises_on_old_server(self, client: weaviate.WeaviateClient) -> None:
+        if client._connection._weaviate_version.is_at_least(1, 37, 0):
+            pytest.skip("Version gate only applies to Weaviate < 1.37.0")
+        with pytest.raises(WeaviateUnsupportedFeatureError):
+            client.tokenization.for_property(
+                collection_name="Any", property_name="title", text="hello"
+            )
+
+
 # ---------------------------------------------------------------------------
 # Async client
 # ---------------------------------------------------------------------------
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index 9ddf5f7ed..de3f68061 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -14,6 +14,7 @@
 )
 from weaviate.connect import executor
 from weaviate.connect.v4 import ConnectionType, _ExpectedStatusCodes
+from weaviate.exceptions import WeaviateUnsupportedFeatureError
 from weaviate.tokenization.models import TokenizeResult
 
 
@@ -55,6 +56,14 @@ class _TokenizationExecutor(Generic[ConnectionType]):
     def __init__(self, connection: ConnectionType):
         self._connection = connection
 
+    def _check_version(self) -> None:
+        if self._connection._weaviate_version.is_lower_than(1, 37, 0):
+            raise WeaviateUnsupportedFeatureError(
+                "Tokenization",
+                str(self._connection._weaviate_version),
+                "1.37.0",
+            )
+
     def text(
         self,
         text: str,
@@ -87,7 +96,11 @@ def text(
 
         Returns:
             A TokenizeResult with indexed and query token lists.
+
+        Raises:
+            WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
         """
+        self._check_version()
         tokenization_str = (
             tokenization.value if isinstance(tokenization, Tokenization) else tokenization
         )
@@ -148,7 +161,11 @@ def for_property(
 
         Returns:
             A TokenizeResult with indexed and query token lists.
+
+        Raises:
+            WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
         """
+        self._check_version()
         path = f"/schema/{collection_name}/properties/{property_name}/tokenize"
 
         payload: Dict[str, Any] = {"text": text}

From 8d379f4142222f7483d29042a9ec75f92c2e4cc6 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Wed, 15 Apr 2026 10:50:22 +0100
Subject: [PATCH 75/99] refactor: update tokenization type to use Tokenization
 enum in TokenizeResult and related tests

---
 integration/test_tokenize.py      | 10 +++++-----
 weaviate/tokenization/executor.py |  2 +-
 weaviate/tokenization/models.py   |  4 ++--
 3 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index 47321aaf5..24b515f2c 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -72,13 +72,13 @@ def test_tokenization_enum(
     ) -> None:
         result = client.tokenization.text(text=text, tokenization=tokenization)
         assert isinstance(result, TokenizeResult)
-        assert result.tokenization == tokenization.value
+        assert result.tokenization == tokenization
         assert result.indexed == expected_tokens
         assert result.query == expected_tokens
 
     def test_tokenization_string(self, client: weaviate.WeaviateClient) -> None:
         result = client.tokenization.text(text="hello world", tokenization="word")
-        assert result.tokenization == "word"
+        assert result.tokenization == Tokenization.WORD
         assert result.indexed == ["hello", "world"]
 
     def test_stopword_preset_enum(self, client: weaviate.WeaviateClient) -> None:
@@ -221,7 +221,7 @@ def test_stopword_config_deserialized_on_property(
                 text="the quick",
             )
             assert isinstance(result, TokenizeResult)
-            assert result.tokenization == "word"
+            assert result.tokenization == Tokenization.WORD
             # Stopword config should be deserialized when present
             if result.stopword_config is not None:
                 assert isinstance(result.stopword_config, StopwordsConfig)
@@ -250,7 +250,7 @@ def test_property_result_types(self, client: weaviate.WeaviateClient) -> None:
                 text="  Hello World  ",
             )
             assert isinstance(result, TokenizeResult)
-            assert result.tokenization == "field"
+            assert result.tokenization == Tokenization.FIELD
             assert result.indexed == ["Hello World"]
         finally:
             client.collections.delete("TestDeserPropTypes")
@@ -371,7 +371,7 @@ async def test_property_tokenize(self, async_client: weaviate.WeaviateAsyncClien
                 text="The quick brown fox",
             )
             assert isinstance(result, TokenizeResult)
-            assert result.tokenization == "word"
+            assert result.tokenization == Tokenization.WORD
             assert result.indexed == ["the", "quick", "brown", "fox"]
             assert "the" not in result.query
             assert "quick" in result.query
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index de3f68061..6228d7350 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -44,7 +44,7 @@ def _parse_stopword_config(body: Dict[str, Any]) -> Optional[StopwordsConfig]:
 
 def _parse_tokenize_result(body: Dict[str, Any]) -> TokenizeResult:
     return TokenizeResult(
-        tokenization=body["tokenization"],
+        tokenization=Tokenization(body["tokenization"]),
         indexed=body["indexed"],
         query=body["query"],
         analyzer_config=_parse_analyzer_config(body),
diff --git a/weaviate/tokenization/models.py b/weaviate/tokenization/models.py
index ba4009b2d..ecb01f695 100644
--- a/weaviate/tokenization/models.py
+++ b/weaviate/tokenization/models.py
@@ -3,7 +3,7 @@
 from dataclasses import dataclass, field
 from typing import List, Optional
 
-from weaviate.collections.classes.config import StopwordsConfig, TextAnalyzerConfig
+from weaviate.collections.classes.config import StopwordsConfig, TextAnalyzerConfig, Tokenization
 
 
 @dataclass
@@ -18,7 +18,7 @@ class TokenizeResult:
         stopword_config: The stopword configuration that was used, if any.
     """
 
-    tokenization: str
+    tokenization: Tokenization
     indexed: List[str]
     query: List[str]
     analyzer_config: Optional[TextAnalyzerConfig] = field(default=None)

From 91a359a38a56b7b812997b3e8280be6ae1d7b71e Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Wed, 15 Apr 2026 11:04:13 +0100
Subject: [PATCH 76/99] refactor: models

---
 integration/test_tokenize.py      | 91 +++++++++++++++----------------
 weaviate/outputs/__init__.py      | 15 ++++-
 weaviate/outputs/tokenization.py  |  5 ++
 weaviate/tokenization/executor.py | 78 ++------------------------
 weaviate/tokenization/models.py   | 47 +++++++++++++---
 5 files changed, 108 insertions(+), 128 deletions(-)
 create mode 100644 weaviate/outputs/tokenization.py

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index 24b515f2c..ddd67b656 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -1,9 +1,10 @@
-"""Integration tests for the tokenize module.
+"""Integration tests for the tokenization module.
 
 These tests cover the client's responsibilities:
-- Correct serialization of inputs (enums, _TextAnalyzerConfigCreate, kwargs)
+- Correct serialization of inputs (enums, _TextAnalyzerConfigCreate, _StopwordsCreate)
 - Correct deserialization of responses into typed objects
 - Client-side validation (_TextAnalyzerConfigCreate rejects invalid input)
+- Version gate (>= 1.37.0)
 - Both sync and async client paths
 """
 
@@ -46,7 +47,7 @@ async def async_client() -> AsyncGenerator[weaviate.WeaviateAsyncClient, None]:
 
 
 # ---------------------------------------------------------------------------
-# Serialization: enums, strings, kwargs, _TextAnalyzerConfigCreate
+# Serialization
 # ---------------------------------------------------------------------------
 
 
@@ -76,72 +77,68 @@ def test_tokenization_enum(
         assert result.indexed == expected_tokens
         assert result.query == expected_tokens
 
-    def test_tokenization_string(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenization.text(text="hello world", tokenization="word")
+    def test_no_analyzer_config(self, client: weaviate.WeaviateClient) -> None:
+        result = client.tokenization.text(text="hello world", tokenization=Tokenization.WORD)
         assert result.tokenization == Tokenization.WORD
         assert result.indexed == ["hello", "world"]
+        assert result.analyzer_config is None
 
-    def test_stopword_preset_enum(self, client: weaviate.WeaviateClient) -> None:
+    def test_ascii_fold(self, client: weaviate.WeaviateClient) -> None:
+        cfg = _TextAnalyzerConfigCreate(ascii_fold=True)
         result = client.tokenization.text(
-            text="The quick brown fox",
+            text="L'école est fermée",
             tokenization=Tokenization.WORD,
-            stopword_preset=StopwordsPreset.EN,
+            analyzer_config=cfg,
         )
-        assert "the" not in result.query
-        assert "quick" in result.query
+        assert result.indexed == ["l", "ecole", "est", "fermee"]
 
-    def test_stopword_preset_string(self, client: weaviate.WeaviateClient) -> None:
+    def test_ascii_fold_with_ignore(self, client: weaviate.WeaviateClient) -> None:
+        cfg = _TextAnalyzerConfigCreate(ascii_fold=True, ascii_fold_ignore=["é"])
         result = client.tokenization.text(
-            text="The quick brown fox",
+            text="L'école est fermée",
             tokenization=Tokenization.WORD,
-            stopword_preset="en",
+            analyzer_config=cfg,
         )
-        assert "the" not in result.query
+        assert result.indexed == ["l", "école", "est", "fermée"]
 
-    def test_ascii_fold_via_kwargs(self, client: weaviate.WeaviateClient) -> None:
+    def test_stopword_preset_enum(self, client: weaviate.WeaviateClient) -> None:
+        cfg = _TextAnalyzerConfigCreate(stopword_preset=StopwordsPreset.EN)
         result = client.tokenization.text(
-            text="L'école est fermée",
+            text="The quick brown fox",
             tokenization=Tokenization.WORD,
-            ascii_fold=True,
+            analyzer_config=cfg,
         )
-        assert result.indexed == ["l", "ecole", "est", "fermee"]
+        assert "the" not in result.query
+        assert "quick" in result.query
 
-    def test_ascii_fold_via_analyzer_config(self, client: weaviate.WeaviateClient) -> None:
-        cfg = _TextAnalyzerConfigCreate(ascii_fold=True)
+    def test_stopword_preset_string(self, client: weaviate.WeaviateClient) -> None:
+        cfg = _TextAnalyzerConfigCreate(stopword_preset="en")
         result = client.tokenization.text(
-            text="L'école est fermée",
+            text="The quick brown fox",
             tokenization=Tokenization.WORD,
             analyzer_config=cfg,
         )
-        assert result.indexed == ["l", "ecole", "est", "fermee"]
+        assert "the" not in result.query
 
-    def test_analyzer_config_and_kwargs_produce_same_result(
-        self, client: weaviate.WeaviateClient
-    ) -> None:
-        """analyzer_config object and equivalent kwargs must produce identical output."""
+    def test_ascii_fold_combined_with_stopwords(self, client: weaviate.WeaviateClient) -> None:
         cfg = _TextAnalyzerConfigCreate(
             ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset=StopwordsPreset.EN
         )
-        via_config = client.tokenization.text(
-            text="L'école est fermée",
+        result = client.tokenization.text(
+            text="The école est fermée",
             tokenization=Tokenization.WORD,
             analyzer_config=cfg,
         )
-        via_kwargs = client.tokenization.text(
-            text="L'école est fermée",
-            tokenization=Tokenization.WORD,
-            ascii_fold=True,
-            ascii_fold_ignore=["é"],
-            stopword_preset=StopwordsPreset.EN,
-        )
-        assert via_config.indexed == via_kwargs.indexed
-        assert via_config.query == via_kwargs.query
+        assert result.indexed == ["the", "école", "est", "fermée"]
+        assert "the" not in result.query
+        assert "école" in result.query
 
-    def test_stopword_presets_serialization(self, client: weaviate.WeaviateClient) -> None:
+    def test_stopword_presets_custom_additions(self, client: weaviate.WeaviateClient) -> None:
+        cfg = _TextAnalyzerConfigCreate(stopword_preset="custom")
         result = client.tokenization.text(
             text="hello world test",
             tokenization=Tokenization.WORD,
-            stopword_preset="custom",
+            analyzer_config=cfg,
             stopword_presets={
                 "custom": _StopwordsCreate(preset=None, additions=["test"], removals=None),
             },
@@ -150,10 +147,11 @@ def test_stopword_presets_serialization(self, client: weaviate.WeaviateClient) -
         assert result.query == ["hello", "world"]
 
     def test_stopword_presets_with_base_and_removals(self, client: weaviate.WeaviateClient) -> None:
+        cfg = _TextAnalyzerConfigCreate(stopword_preset="en-no-the")
         result = client.tokenization.text(
             text="the quick",
             tokenization=Tokenization.WORD,
-            stopword_preset="en-no-the",
+            analyzer_config=cfg,
             stopword_presets={
                 "en-no-the": _StopwordsCreate(
                     preset=StopwordsPreset.EN, additions=None, removals=["the"]
@@ -165,7 +163,7 @@ def test_stopword_presets_with_base_and_removals(self, client: weaviate.Weaviate
 
 
 # ---------------------------------------------------------------------------
-# Deserialization: typed response fields
+# Deserialization
 # ---------------------------------------------------------------------------
 
 
@@ -179,12 +177,13 @@ def test_result_type(self, client: weaviate.WeaviateClient) -> None:
         assert isinstance(result.query, list)
 
     def test_analyzer_config_deserialized(self, client: weaviate.WeaviateClient) -> None:
+        cfg = _TextAnalyzerConfigCreate(
+            ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset=StopwordsPreset.EN
+        )
         result = client.tokenization.text(
             text="L'école",
             tokenization=Tokenization.WORD,
-            ascii_fold=True,
-            ascii_fold_ignore=["é"],
-            stopword_preset=StopwordsPreset.EN,
+            analyzer_config=cfg,
         )
         assert isinstance(result.analyzer_config, TextAnalyzerConfig)
         assert result.analyzer_config.ascii_fold is True
@@ -198,7 +197,6 @@ def test_no_analyzer_config_returns_none(self, client: weaviate.WeaviateClient)
     def test_stopword_config_deserialized_on_property(
         self, client: weaviate.WeaviateClient
     ) -> None:
-        """Property endpoint returns stopwordConfig; verify it deserializes to StopwordsConfig."""
         client.collections.delete("TestDeserStopword")
         try:
             client.collections.create_from_dict(
@@ -222,7 +220,6 @@ def test_stopword_config_deserialized_on_property(
             )
             assert isinstance(result, TokenizeResult)
             assert result.tokenization == Tokenization.WORD
-            # Stopword config should be deserialized when present
             if result.stopword_config is not None:
                 assert isinstance(result.stopword_config, StopwordsConfig)
         finally:
@@ -322,7 +319,7 @@ def test_for_property_raises_on_old_server(self, client: weaviate.WeaviateClient
 
 
 class TestAsyncClient:
-    """Verify both text() and property() work through the async client."""
+    """Verify both text() and for_property() work through the async client."""
 
     @pytest.mark.asyncio
     async def test_text_tokenize(self, async_client: weaviate.WeaviateAsyncClient) -> None:
diff --git a/weaviate/outputs/__init__.py b/weaviate/outputs/__init__.py
index 62193fc35..ba3cf894f 100644
--- a/weaviate/outputs/__init__.py
+++ b/weaviate/outputs/__init__.py
@@ -1,4 +1,16 @@
-from . import aggregate, backup, batch, cluster, config, data, query, replication, tenants, users
+from . import (
+    aggregate,
+    backup,
+    batch,
+    cluster,
+    config,
+    data,
+    query,
+    replication,
+    tenants,
+    tokenization,
+    users,
+)
 
 __all__ = [
     "aggregate",
@@ -10,5 +22,6 @@
     "query",
     "replication",
     "tenants",
+    "tokenization",
     "users",
 ]
diff --git a/weaviate/outputs/tokenization.py b/weaviate/outputs/tokenization.py
new file mode 100644
index 000000000..0854f8b0d
--- /dev/null
+++ b/weaviate/outputs/tokenization.py
@@ -0,0 +1,5 @@
+from weaviate.tokenization.models import TokenizeResult
+
+__all__ = [
+    "TokenizeResult",
+]
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index 6228d7350..226aeb6c6 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -1,13 +1,10 @@
 """Tokenize executor."""
 
-from typing import Any, Dict, Generic, List, Optional, Union
+from typing import Any, Dict, Generic, Optional
 
 from httpx import Response
 
 from weaviate.collections.classes.config import (
-    StopwordsConfig,
-    StopwordsPreset,
-    TextAnalyzerConfig,
     Tokenization,
     _StopwordsCreate,
     _TextAnalyzerConfigCreate,
@@ -18,40 +15,6 @@
 from weaviate.tokenization.models import TokenizeResult
 
 
-def _parse_analyzer_config(body: Dict[str, Any]) -> Optional[TextAnalyzerConfig]:
-    ac = body.get("analyzerConfig")
-    if ac is None:
-        return None
-    if "asciiFold" not in ac and "stopwordPreset" not in ac:
-        return None
-    return TextAnalyzerConfig(
-        ascii_fold=ac.get("asciiFold", False),
-        ascii_fold_ignore=ac.get("asciiFoldIgnore"),
-        stopword_preset=ac.get("stopwordPreset"),
-    )
-
-
-def _parse_stopword_config(body: Dict[str, Any]) -> Optional[StopwordsConfig]:
-    sc = body.get("stopwordConfig")
-    if sc is None:
-        return None
-    return StopwordsConfig(
-        preset=StopwordsPreset(sc["preset"]) if sc.get("preset") else StopwordsPreset.NONE,
-        additions=sc.get("additions"),
-        removals=sc.get("removals"),
-    )
-
-
-def _parse_tokenize_result(body: Dict[str, Any]) -> TokenizeResult:
-    return TokenizeResult(
-        tokenization=Tokenization(body["tokenization"]),
-        indexed=body["indexed"],
-        query=body["query"],
-        analyzer_config=_parse_analyzer_config(body),
-        stopword_config=_parse_stopword_config(body),
-    )
-
-
 class _TokenizationExecutor(Generic[ConnectionType]):
     def __init__(self, connection: ConnectionType):
         self._connection = connection
@@ -67,29 +30,17 @@ def _check_version(self) -> None:
     def text(
         self,
         text: str,
-        tokenization: Union[Tokenization, str],
+        tokenization: Tokenization,
         *,
         analyzer_config: Optional[_TextAnalyzerConfigCreate] = None,
-        ascii_fold: Optional[bool] = None,
-        ascii_fold_ignore: Optional[List[str]] = None,
-        stopword_preset: Optional[Union[StopwordsPreset, str]] = None,
         stopword_presets: Optional[Dict[str, _StopwordsCreate]] = None,
     ) -> executor.Result[TokenizeResult]:
         """Tokenize text using the generic /v1/tokenize endpoint.
 
-        Analyzer settings can be provided either via a ``_TextAnalyzerConfigCreate``
-        object **or** via the individual keyword arguments (``ascii_fold``,
-        ``ascii_fold_ignore``, ``stopword_preset``).  If ``analyzer_config`` is
-        given the individual keyword arguments are ignored.
-
         Args:
             text: The text to tokenize.
             tokenization: The tokenization method to use (e.g. Tokenization.WORD).
-            analyzer_config: A ``_TextAnalyzerConfigCreate`` instance that bundles
-                ascii_fold, ascii_fold_ignore, and stopword_preset settings.
-            ascii_fold: Whether to fold accented characters to ASCII equivalents.
-            ascii_fold_ignore: Characters to exclude from ASCII folding.
-            stopword_preset: Stopword preset name to apply for query-time filtering.
+            analyzer_config: Text analyzer settings (ASCII folding, stopword preset).
             stopword_presets: Custom stopword preset definitions, keyed by name.
                 Each value is a ``_StopwordsCreate`` with optional preset, additions,
                 and removals fields.
@@ -101,33 +52,16 @@ def text(
             WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
         """
         self._check_version()
-        tokenization_str = (
-            tokenization.value if isinstance(tokenization, Tokenization) else tokenization
-        )
 
         payload: Dict[str, Any] = {
             "text": text,
-            "tokenization": tokenization_str,
+            "tokenization": tokenization.value,
         }
 
         if analyzer_config is not None:
             ac_dict = analyzer_config._to_dict()
             if ac_dict:
                 payload["analyzerConfig"] = ac_dict
-        else:
-            ac: Dict[str, Any] = {}
-            if ascii_fold is not None:
-                ac["asciiFold"] = ascii_fold
-            if ascii_fold_ignore is not None:
-                ac["asciiFoldIgnore"] = ascii_fold_ignore
-            if stopword_preset is not None:
-                ac["stopwordPreset"] = (
-                    stopword_preset.value
-                    if isinstance(stopword_preset, StopwordsPreset)
-                    else stopword_preset
-                )
-            if ac:
-                payload["analyzerConfig"] = ac
 
         if stopword_presets is not None:
             payload["stopwordPresets"] = {
@@ -135,7 +69,7 @@ def text(
             }
 
         def resp(response: Response) -> TokenizeResult:
-            return _parse_tokenize_result(response.json())
+            return TokenizeResult.model_validate(response.json())
 
         return executor.execute(
             response_callback=resp,
@@ -171,7 +105,7 @@ def for_property(
         payload: Dict[str, Any] = {"text": text}
 
         def resp(response: Response) -> TokenizeResult:
-            return _parse_tokenize_result(response.json())
+            return TokenizeResult.model_validate(response.json())
 
         return executor.execute(
             response_callback=resp,
diff --git a/weaviate/tokenization/models.py b/weaviate/tokenization/models.py
index ecb01f695..8bfa508f8 100644
--- a/weaviate/tokenization/models.py
+++ b/weaviate/tokenization/models.py
@@ -1,13 +1,18 @@
-"""Return types for tokenize operations."""
+"""Return types for tokenization operations."""
 
-from dataclasses import dataclass, field
-from typing import List, Optional
+from typing import Any, Dict, List, Optional
 
-from weaviate.collections.classes.config import StopwordsConfig, TextAnalyzerConfig, Tokenization
+from pydantic import BaseModel, ConfigDict, Field, field_validator
 
+from weaviate.collections.classes.config import (
+    StopwordsConfig,
+    StopwordsPreset,
+    TextAnalyzerConfig,
+    Tokenization,
+)
 
-@dataclass
-class TokenizeResult:
+
+class TokenizeResult(BaseModel):
     """Result of a tokenization operation.
 
     Attributes:
@@ -18,8 +23,34 @@ class TokenizeResult:
         stopword_config: The stopword configuration that was used, if any.
     """
 
+    model_config = ConfigDict(populate_by_name=True, arbitrary_types_allowed=True)
+
     tokenization: Tokenization
     indexed: List[str]
     query: List[str]
-    analyzer_config: Optional[TextAnalyzerConfig] = field(default=None)
-    stopword_config: Optional[StopwordsConfig] = field(default=None)
+    analyzer_config: Optional[TextAnalyzerConfig] = Field(default=None, alias="analyzerConfig")
+    stopword_config: Optional[StopwordsConfig] = Field(default=None, alias="stopwordConfig")
+
+    @field_validator("analyzer_config", mode="before")
+    @classmethod
+    def _parse_analyzer_config(cls, v: Optional[Dict[str, Any]]) -> Optional[TextAnalyzerConfig]:
+        if v is None:
+            return None
+        if "asciiFold" not in v and "stopwordPreset" not in v:
+            return None
+        return TextAnalyzerConfig(
+            ascii_fold=v.get("asciiFold", False),
+            ascii_fold_ignore=v.get("asciiFoldIgnore"),
+            stopword_preset=v.get("stopwordPreset"),
+        )
+
+    @field_validator("stopword_config", mode="before")
+    @classmethod
+    def _parse_stopword_config(cls, v: Optional[Dict[str, Any]]) -> Optional[StopwordsConfig]:
+        if v is None:
+            return None
+        return StopwordsConfig(
+            preset=StopwordsPreset(v["preset"]),
+            additions=v.get("additions"),
+            removals=v.get("removals"),
+        )

From 61665e712dac3d6e0665b9e5e8e7ae85d8e47144 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Wed, 15 Apr 2026 11:08:31 +0100
Subject: [PATCH 77/99] refactor: move tokenize property to class config

---
 integration/test_tokenize.py            | 27 +++++++----------
 weaviate/collections/config/executor.py | 40 +++++++++++++++++++++++++
 weaviate/tokenization/executor.py       | 36 ----------------------
 3 files changed, 50 insertions(+), 53 deletions(-)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index ddd67b656..565cb197d 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -213,11 +213,8 @@ def test_stopword_config_deserialized_on_property(
                     ],
                 }
             )
-            result = client.tokenization.for_property(
-                collection_name="TestDeserStopword",
-                property_name="title",
-                text="the quick",
-            )
+            col = client.collections.get("TestDeserStopword")
+            result = col.config.tokenize_property(property_name="title", text="the quick")
             assert isinstance(result, TokenizeResult)
             assert result.tokenization == Tokenization.WORD
             if result.stopword_config is not None:
@@ -241,11 +238,8 @@ def test_property_result_types(self, client: weaviate.WeaviateClient) -> None:
                     ],
                 }
             )
-            result = client.tokenization.for_property(
-                collection_name="TestDeserPropTypes",
-                property_name="tag",
-                text="  Hello World  ",
-            )
+            col = client.collections.get("TestDeserPropTypes")
+            result = col.config.tokenize_property(property_name="tag", text="  Hello World  ")
             assert isinstance(result, TokenizeResult)
             assert result.tokenization == Tokenization.FIELD
             assert result.indexed == ["Hello World"]
@@ -304,13 +298,12 @@ def test_text_raises_on_old_server(self, client: weaviate.WeaviateClient) -> Non
         with pytest.raises(WeaviateUnsupportedFeatureError):
             client.tokenization.text(text="hello", tokenization=Tokenization.WORD)
 
-    def test_for_property_raises_on_old_server(self, client: weaviate.WeaviateClient) -> None:
+    def test_tokenize_property_raises_on_old_server(self, client: weaviate.WeaviateClient) -> None:
         if client._connection._weaviate_version.is_at_least(1, 37, 0):
             pytest.skip("Version gate only applies to Weaviate < 1.37.0")
+        col = client.collections.get("Any")
         with pytest.raises(WeaviateUnsupportedFeatureError):
-            client.tokenization.for_property(
-                collection_name="Any", property_name="title", text="hello"
-            )
+            col.config.tokenize_property(property_name="title", text="hello")
 
 
 # ---------------------------------------------------------------------------
@@ -319,7 +312,7 @@ def test_for_property_raises_on_old_server(self, client: weaviate.WeaviateClient
 
 
 class TestAsyncClient:
-    """Verify both text() and for_property() work through the async client."""
+    """Verify text() and tokenize_property() work through the async client."""
 
     @pytest.mark.asyncio
     async def test_text_tokenize(self, async_client: weaviate.WeaviateAsyncClient) -> None:
@@ -362,8 +355,8 @@ async def test_property_tokenize(self, async_client: weaviate.WeaviateAsyncClien
                     ],
                 }
             )
-            result = await async_client.tokenization.for_property(
-                collection_name="TestAsyncPropTokenize",
+            col = async_client.collections.get("TestAsyncPropTokenize")
+            result = await col.config.tokenize_property(
                 property_name="title",
                 text="The quick brown fox",
             )
diff --git a/weaviate/collections/config/executor.py b/weaviate/collections/config/executor.py
index bd302c3c4..9e9390cda 100644
--- a/weaviate/collections/config/executor.py
+++ b/weaviate/collections/config/executor.py
@@ -56,6 +56,7 @@
     WeaviateInvalidInputError,
     WeaviateUnsupportedFeatureError,
 )
+from weaviate.tokenization.models import TokenizeResult
 from weaviate.util import (
     _capitalize_first_letter,
     _decode_json_response_dict,
@@ -666,3 +667,42 @@ def resp(res: Response) -> bool:
             error_msg="Property may not exist",
             status_codes=_ExpectedStatusCodes(ok_in=[200], error="property exists"),
         )
+
+    def tokenize_property(
+        self,
+        property_name: str,
+        text: str,
+    ) -> executor.Result[TokenizeResult]:
+        """Tokenize text using a property's configured tokenization settings.
+
+        Args:
+            property_name: The property name whose tokenization config to use.
+            text: The text to tokenize.
+
+        Returns:
+            A TokenizeResult with indexed and query token lists.
+
+        Raises:
+            WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
+        """
+        if self._connection._weaviate_version.is_lower_than(1, 37, 0):
+            raise WeaviateUnsupportedFeatureError(
+                "Tokenization",
+                str(self._connection._weaviate_version),
+                "1.37.0",
+            )
+
+        path = f"/schema/{self._name}/properties/{property_name}/tokenize"
+        payload: Dict[str, Any] = {"text": text}
+
+        def resp(response: Response) -> TokenizeResult:
+            return TokenizeResult.model_validate(response.json())
+
+        return executor.execute(
+            response_callback=resp,
+            method=self._connection.post,
+            path=path,
+            weaviate_object=payload,
+            error_msg="Property tokenization failed",
+            status_codes=_ExpectedStatusCodes(ok_in=[200], error="tokenize property text"),
+        )
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index 226aeb6c6..5093c14e9 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -79,39 +79,3 @@ def resp(response: Response) -> TokenizeResult:
             error_msg="Tokenization failed",
             status_codes=_ExpectedStatusCodes(ok_in=[200], error="tokenize text"),
         )
-
-    def for_property(
-        self,
-        collection_name: str,
-        property_name: str,
-        text: str,
-    ) -> executor.Result[TokenizeResult]:
-        """Tokenize text using a property's configured tokenization settings.
-
-        Args:
-            collection_name: The collection (class) name.
-            property_name: The property name whose tokenization config to use.
-            text: The text to tokenize.
-
-        Returns:
-            A TokenizeResult with indexed and query token lists.
-
-        Raises:
-            WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
-        """
-        self._check_version()
-        path = f"/schema/{collection_name}/properties/{property_name}/tokenize"
-
-        payload: Dict[str, Any] = {"text": text}
-
-        def resp(response: Response) -> TokenizeResult:
-            return TokenizeResult.model_validate(response.json())
-
-        return executor.execute(
-            response_callback=resp,
-            method=self._connection.post,
-            path=path,
-            weaviate_object=payload,
-            error_msg="Property tokenization failed",
-            status_codes=_ExpectedStatusCodes(ok_in=[200], error="tokenize property text"),
-        )

From aea03278f3ee5712608589138637d77131364955 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Wed, 15 Apr 2026 11:16:24 +0100
Subject: [PATCH 78/99] fix: remove trailing whitespace in __init__.py

---
 weaviate/outputs/__init__.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/weaviate/outputs/__init__.py b/weaviate/outputs/__init__.py
index 5381d6dc4..75cb031e0 100644
--- a/weaviate/outputs/__init__.py
+++ b/weaviate/outputs/__init__.py
@@ -9,7 +9,7 @@
     query,
     replication,
     tenants,
-    tokenization,  
+    tokenization,
     users,
 )
 

From ef55ce283b1aae518cb8eacc6dba7fdf6530b709 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Wed, 15 Apr 2026 11:55:39 +0100
Subject: [PATCH 79/99] test: add version gate for Weaviate >= 1.37.0 in
 tokenization tests

---
 integration/test_tokenize.py | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index 565cb197d..97587235b 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -36,6 +36,12 @@ def client() -> Generator[weaviate.WeaviateClient, None, None]:
     c.close()
 
 
+@pytest.fixture(autouse=False)
+def require_1_37(client: weaviate.WeaviateClient) -> None:
+    if client._connection._weaviate_version.is_lower_than(1, 37, 0):
+        pytest.skip("Tokenization requires Weaviate >= 1.37.0")
+
+
 @pytest_asyncio.fixture
 async def async_client() -> AsyncGenerator[weaviate.WeaviateAsyncClient, None]:
     c = weaviate.use_async_with_local(
@@ -51,6 +57,7 @@ async def async_client() -> AsyncGenerator[weaviate.WeaviateAsyncClient, None]:
 # ---------------------------------------------------------------------------
 
 
+@pytest.mark.usefixtures("require_1_37")
 class TestSerialization:
     """Verify the client correctly serializes different input forms."""
 
@@ -167,6 +174,7 @@ def test_stopword_presets_with_base_and_removals(self, client: weaviate.Weaviate
 # ---------------------------------------------------------------------------
 
 
+@pytest.mark.usefixtures("require_1_37")
 class TestDeserialization:
     """Verify the client correctly deserializes response fields into typed objects."""
 
@@ -311,6 +319,7 @@ def test_tokenize_property_raises_on_old_server(self, client: weaviate.WeaviateC
 # ---------------------------------------------------------------------------
 
 
+@pytest.mark.usefixtures("require_1_37")
 class TestAsyncClient:
     """Verify text() and tokenize_property() work through the async client."""
 

From dff05f534c200662f7bf820cecb4efb77a5957ac Mon Sep 17 00:00:00 2001
From: Marcin Antas <antas.marcin@gmail.com>
Date: Sat, 14 Mar 2026 12:03:38 +0100
Subject: [PATCH 80/99] feat: add support for blobHash property type

---
 test/collection/test_config.py         | 8 ++++++++
 weaviate/collections/classes/config.py | 2 ++
 2 files changed, 10 insertions(+)

diff --git a/test/collection/test_config.py b/test/collection/test_config.py
index 5a69b21c5..84bba4a63 100644
--- a/test/collection/test_config.py
+++ b/test/collection/test_config.py
@@ -1340,6 +1340,10 @@ def test_config_create_with_properties(
                 name="blob",
                 data_type=DataType.BLOB,
             ),
+            Property(
+                name="blob_hash",
+                data_type=DataType.BLOB_HASH,
+            ),
             Property(
                 name="phone_number",
                 data_type=DataType.PHONE_NUMBER,
@@ -1403,6 +1407,10 @@ def test_config_create_with_properties(
             "dataType": ["blob"],
             "name": "blob",
         },
+        {
+            "dataType": ["blobHash"],
+            "name": "blob_hash",
+        },
         {
             "dataType": ["phoneNumber"],
             "name": "phone_number",
diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 7874cd8e2..6d60482a3 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -148,6 +148,7 @@ class DataType(str, BaseEnum):
         UUID_ARRAY: UUID array data type.
         GEO_COORDINATES: Geo coordinates data type.
         BLOB: Blob data type.
+        BLOB_HASH: Blob hash data type.
         PHONE_NUMBER: Phone number data type.
         OBJECT: Object data type.
         OBJECT_ARRAY: Object array data type.
@@ -167,6 +168,7 @@ class DataType(str, BaseEnum):
     UUID_ARRAY = "uuid[]"
     GEO_COORDINATES = "geoCoordinates"
     BLOB = "blob"
+    BLOB_HASH = "blobHash"
     PHONE_NUMBER = "phoneNumber"
     OBJECT = "object"
     OBJECT_ARRAY = "object[]"

From 906b35b62441aeca41e7628d6ddeace5dbc30bb2 Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Thu, 16 Apr 2026 16:00:25 +0200
Subject: [PATCH 81/99] Add full_with_profile

---
 integration/test_collection_query_profile.py | 30 ++++++++++++++++++++
 weaviate/collections/classes/grpc.py         | 24 +++++++++++++++-
 2 files changed, 53 insertions(+), 1 deletion(-)

diff --git a/integration/test_collection_query_profile.py b/integration/test_collection_query_profile.py
index 0c68c9e69..da93b90f5 100644
--- a/integration/test_collection_query_profile.py
+++ b/integration/test_collection_query_profile.py
@@ -173,6 +173,36 @@ def test_near_vector_group_by_with_query_profile(
     assert_common_profile(shard.searches["vector"])
 
 
+def test_full_with_profile(collection_factory: CollectionFactory) -> None:
+    """Test that MetadataQuery.full_with_profile() returns profiling and all other metadata."""
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.near_vector(
+        near_vector=[1.0, 0.0, 0.0],
+        return_metadata=MetadataQuery.full_with_profile(),
+        limit=1,
+    )
+    assert len(result.objects) == 1
+    obj = result.objects[0]
+    assert obj.metadata.distance is not None
+    assert obj.metadata.creation_time is not None
+    assert obj.metadata.last_update_time is not None
+    assert obj.metadata.score is not None
+    assert obj.metadata.explain_score is not None
+
+    assert result.query_profile is not None
+    assert len(result.query_profile.shards) > 0
+    assert_common_profile(result.query_profile.shards[0].searches["vector"])
+
+
+def test_full_excludes_query_profile(collection_factory: CollectionFactory) -> None:
+    """Test that MetadataQuery.full() does not include query profiling."""
+    collection = _create_and_populate(collection_factory)
+    result = collection.query.fetch_objects(
+        return_metadata=MetadataQuery.full(),
+    )
+    assert result.query_profile is None
+
+
 def test_no_query_profile_when_not_requested(
     collection_factory: CollectionFactory,
 ) -> None:
diff --git a/weaviate/collections/classes/grpc.py b/weaviate/collections/classes/grpc.py
index 20ece3bc0..bdcc53dd9 100644
--- a/weaviate/collections/classes/grpc.py
+++ b/weaviate/collections/classes/grpc.py
@@ -94,7 +94,28 @@ class MetadataQuery(_WeaviateInput):
 
     @classmethod
     def full(cls) -> "MetadataQuery":
-        """Return a MetadataQuery with all fields set to True."""
+        """Return a MetadataQuery with all fields set to True.
+
+        NOTE: `query_profile` is excluded because it adds performance overhead.
+        Use `full_with_profile()` to include it.
+        """
+        return cls(
+            creation_time=True,
+            last_update_time=True,
+            distance=True,
+            certainty=True,
+            score=True,
+            explain_score=True,
+            is_consistent=True,
+        )
+
+    @classmethod
+    def full_with_profile(cls) -> "MetadataQuery":
+        """Return a MetadataQuery with all fields set to True, including query profiling.
+
+        Query profiling adds per-shard execution timing breakdowns to the response
+        but has performance overhead. Requires Weaviate >= 1.36.9.
+        """
         return cls(
             creation_time=True,
             last_update_time=True,
@@ -103,6 +124,7 @@ def full(cls) -> "MetadataQuery":
             score=True,
             explain_score=True,
             is_consistent=True,
+            query_profile=True,
         )
 
 

From 66a2fb2e3e164c13e9251ae896e1adc431f9d971 Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Fri, 17 Apr 2026 12:51:19 +0200
Subject: [PATCH 82/99] Refactor RBAC permissions

---
 integration/test_rbac.py | 29 ++++++++++++++++++++++++++---
 weaviate/rbac/models.py  | 19 +++++++++++++++----
 2 files changed, 41 insertions(+), 7 deletions(-)

diff --git a/integration/test_rbac.py b/integration/test_rbac.py
index 86719b6ce..0f8657a2d 100644
--- a/integration/test_rbac.py
+++ b/integration/test_rbac.py
@@ -427,9 +427,9 @@
             32,  # Minimum version for alias permissions
         ),
         (
-            Permissions.mcp(manage=True),
+            Permissions.mcp(create=True, read=True, update=True),
             Role(
-                name="ManageMCP",
+                name="MCPAll",
                 alias_permissions=[],
                 cluster_permissions=[],
                 users_permissions=[],
@@ -437,7 +437,30 @@
                 roles_permissions=[],
                 data_permissions=[],
                 backups_permissions=[],
-                mcp_permissions=[MCPPermissionOutput(actions={Actions.MCP.MANAGE})],
+                mcp_permissions=[
+                    MCPPermissionOutput(
+                        actions={Actions.MCP.CREATE, Actions.MCP.READ, Actions.MCP.UPDATE}
+                    )
+                ],
+                nodes_permissions=[],
+                tenants_permissions=[],
+                replicate_permissions=[],
+                groups_permissions=[],
+            ),
+            37,  # Minimum version for MCP permissions
+        ),
+        (
+            Permissions.mcp(read=True),
+            Role(
+                name="MCPRead",
+                alias_permissions=[],
+                cluster_permissions=[],
+                users_permissions=[],
+                collections_permissions=[],
+                roles_permissions=[],
+                data_permissions=[],
+                backups_permissions=[],
+                mcp_permissions=[MCPPermissionOutput(actions={Actions.MCP.READ})],
                 nodes_permissions=[],
                 tenants_permissions=[],
                 replicate_permissions=[],
diff --git a/weaviate/rbac/models.py b/weaviate/rbac/models.py
index dfdbc48f4..8e0989542 100644
--- a/weaviate/rbac/models.py
+++ b/weaviate/rbac/models.py
@@ -253,7 +253,9 @@ def values() -> List[str]:
 
 
 class MCPAction(str, _Action, Enum):
-    MANAGE = "manage_mcp"
+    CREATE = "create_mcp"
+    READ = "read_mcp"
+    UPDATE = "update_mcp"
 
     @staticmethod
     def values() -> List[str]:
@@ -1053,9 +1055,18 @@ def backup(
         return permissions
 
     @staticmethod
-    def mcp(*, manage: bool = False) -> PermissionsCreateType:
-        if manage:
-            return [_MCPPermission(actions={MCPAction.MANAGE})]
+    def mcp(
+        *, create: bool = False, read: bool = False, update: bool = False
+    ) -> PermissionsCreateType:
+        actions: Set[MCPAction] = set()
+        if create:
+            actions.add(MCPAction.CREATE)
+        if read:
+            actions.add(MCPAction.READ)
+        if update:
+            actions.add(MCPAction.UPDATE)
+        if len(actions) > 0:
+            return [_MCPPermission(actions=actions)]
         return []
 
     @staticmethod

From 0955364bce3c6359737e857ceebfb948a162fb8e Mon Sep 17 00:00:00 2001
From: Ivan Despot <66276597+g-despot@users.noreply.github.com>
Date: Mon, 20 Apr 2026 08:33:08 +0200
Subject: [PATCH 83/99] Bump Weaviate version

---
 .github/workflows/main.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index a7d6cca4d..8dd157443 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -28,7 +28,7 @@ env:
   WEAVIATE_134: 1.34.19
   WEAVIATE_135: 1.35.16-efdedfa
   WEAVIATE_136: 1.36.9-d905e6c
-  WEAVIATE_137: 1.37.0-rc.1-bc3891e
+  WEAVIATE_137: 1.37.1
 
 jobs:
   lint-and-format:

From 3624e8b571d8c388befc61e232e0cecff832b0f0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 20 Apr 2026 15:52:55 +0100
Subject: [PATCH 84/99] refactor: tokenization executor and models to support
 stopword configurations and improve response handling

---
 integration/test_tokenize.py      | 245 +++++++++++++++++++-----------
 weaviate/tokenization/executor.py |  49 +++++-
 weaviate/tokenization/models.py   |  46 +-----
 3 files changed, 202 insertions(+), 138 deletions(-)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index 97587235b..d692a4808 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -2,10 +2,17 @@
 
 These tests cover the client's responsibilities:
 - Correct serialization of inputs (enums, _TextAnalyzerConfigCreate, _StopwordsCreate)
-- Correct deserialization of responses into typed objects
-- Client-side validation (_TextAnalyzerConfigCreate rejects invalid input)
+- Correct deserialization of responses into the TokenizeResult object
+- Client-side validation (_TextAnalyzerConfigCreate, stopwords/stopword_presets mutex)
 - Version gate (>= 1.37.0)
 - Both sync and async client paths
+
+Server-side behavior this client relies on:
+- Word tokenization defaults to preset "en" when no stopword config is sent.
+- The generic /v1/tokenize response is minimal: only ``indexed`` and ``query``
+  are returned. The property-level endpoint additionally returns ``tokenization``.
+- ``stopwords`` and ``stopword_presets`` are mutually exclusive on the generic
+  endpoint — the server rejects requests that set both.
 """
 
 from typing import AsyncGenerator, Generator
@@ -15,9 +22,7 @@
 
 import weaviate
 from weaviate.collections.classes.config import (
-    StopwordsConfig,
     StopwordsPreset,
-    TextAnalyzerConfig,
     Tokenization,
     _StopwordsCreate,
     _TextAnalyzerConfigCreate,
@@ -62,13 +67,31 @@ class TestSerialization:
     """Verify the client correctly serializes different input forms."""
 
     @pytest.mark.parametrize(
-        "tokenization,text,expected_tokens",
+        "tokenization,text,expected_indexed,expected_query",
         [
-            (Tokenization.WORD, "The quick brown fox", ["the", "quick", "brown", "fox"]),
-            (Tokenization.LOWERCASE, "Hello World Test", ["hello", "world", "test"]),
-            (Tokenization.WHITESPACE, "Hello World Test", ["Hello", "World", "Test"]),
-            (Tokenization.FIELD, "  Hello World  ", ["Hello World"]),
-            (Tokenization.TRIGRAM, "Hello", ["hel", "ell", "llo"]),
+            # "the" is an English stopword — filtered from the query output
+            # by the server's default "en" preset for word tokenization.
+            (
+                Tokenization.WORD,
+                "The quick brown fox",
+                ["the", "quick", "brown", "fox"],
+                ["quick", "brown", "fox"],
+            ),
+            # Non-word tokenizations do not apply the default "en" preset.
+            (
+                Tokenization.LOWERCASE,
+                "Hello World Test",
+                ["hello", "world", "test"],
+                ["hello", "world", "test"],
+            ),
+            (
+                Tokenization.WHITESPACE,
+                "Hello World Test",
+                ["Hello", "World", "Test"],
+                ["Hello", "World", "Test"],
+            ),
+            (Tokenization.FIELD, "  Hello World  ", ["Hello World"], ["Hello World"]),
+            (Tokenization.TRIGRAM, "Hello", ["hel", "ell", "llo"], ["hel", "ell", "llo"]),
         ],
     )
     def test_tokenization_enum(
@@ -76,19 +99,35 @@ def test_tokenization_enum(
         client: weaviate.WeaviateClient,
         tokenization: Tokenization,
         text: str,
-        expected_tokens: list,
+        expected_indexed: list,
+        expected_query: list,
     ) -> None:
         result = client.tokenization.text(text=text, tokenization=tokenization)
         assert isinstance(result, TokenizeResult)
-        assert result.tokenization == tokenization
-        assert result.indexed == expected_tokens
-        assert result.query == expected_tokens
+        assert result.indexed == expected_indexed
+        assert result.query == expected_query
+        # Generic endpoint does not echo tokenization back.
+        assert result.tokenization is None
+
+    def test_default_en_applied_for_word(self, client: weaviate.WeaviateClient) -> None:
+        """Word tokenization defaults to the 'en' preset when no stopword
+        config is supplied."""
+        result = client.tokenization.text(
+            text="The quick brown fox", tokenization=Tokenization.WORD
+        )
+        assert result.indexed == ["the", "quick", "brown", "fox"]
+        # "the" removed by the server's default en preset.
+        assert result.query == ["quick", "brown", "fox"]
 
-    def test_no_analyzer_config(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenization.text(text="hello world", tokenization=Tokenization.WORD)
-        assert result.tokenization == Tokenization.WORD
-        assert result.indexed == ["hello", "world"]
-        assert result.analyzer_config is None
+    def test_opt_out_of_default_en(self, client: weaviate.WeaviateClient) -> None:
+        """analyzerConfig.stopwordPreset='none' disables the default en."""
+        cfg = _TextAnalyzerConfigCreate(stopword_preset=StopwordsPreset.NONE)
+        result = client.tokenization.text(
+            text="The quick brown fox",
+            tokenization=Tokenization.WORD,
+            analyzer_config=cfg,
+        )
+        assert result.query == ["the", "quick", "brown", "fox"]
 
     def test_ascii_fold(self, client: weaviate.WeaviateClient) -> None:
         cfg = _TextAnalyzerConfigCreate(ascii_fold=True)
@@ -140,33 +179,74 @@ def test_ascii_fold_combined_with_stopwords(self, client: weaviate.WeaviateClien
         assert "the" not in result.query
         assert "école" in result.query
 
-    def test_stopword_presets_custom_additions(self, client: weaviate.WeaviateClient) -> None:
-        cfg = _TextAnalyzerConfigCreate(stopword_preset="custom")
+    def test_stopwords_fallback(self, client: weaviate.WeaviateClient) -> None:
+        """Top-level stopwords acts as the fallback detector when no
+        analyzerConfig.stopwordPreset is set."""
+        sw = _StopwordsCreate(
+            preset=StopwordsPreset.EN, additions=["quick"], removals=None
+        )
+        result = client.tokenization.text(
+            text="the quick brown fox",
+            tokenization=Tokenization.WORD,
+            stopwords=sw,
+        )
+        assert result.indexed == ["the", "quick", "brown", "fox"]
+        # "the" (en) and "quick" (addition) filtered.
+        assert result.query == ["brown", "fox"]
+
+    def test_stopwords_additions_default_preset_to_en(
+        self, client: weaviate.WeaviateClient
+    ) -> None:
+        """Caller omits preset, passes only additions. Server defaults preset
+        to 'en' and builds detector from en + additions."""
+        sw = _StopwordsCreate(preset=None, additions=["hello"], removals=None)
+        result = client.tokenization.text(
+            text="the quick hello world",
+            tokenization=Tokenization.WORD,
+            stopwords=sw,
+        )
+        assert result.query == ["quick", "world"]
+
+    def test_stopwords_removals_default_preset_to_en(
+        self, client: weaviate.WeaviateClient
+    ) -> None:
+        """Caller omits preset, passes only removals. 'the' is removed from
+        the en list so it passes through."""
+        sw = _StopwordsCreate(preset=None, additions=None, removals=["the"])
+        result = client.tokenization.text(
+            text="the quick is fast",
+            tokenization=Tokenization.WORD,
+            stopwords=sw,
+        )
+        # "is" still in en, "the" removed.
+        assert result.query == ["the", "quick", "fast"]
+
+    def test_stopword_presets_named_reference(self, client: weaviate.WeaviateClient) -> None:
+        """Define a named preset via stopword_presets, select it via
+        analyzerConfig.stopwordPreset. Word lists use the collection shape."""
         result = client.tokenization.text(
             text="hello world test",
             tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
-            stopword_presets={
-                "custom": _StopwordsCreate(preset=None, additions=["test"], removals=None),
-            },
+            analyzer_config=_TextAnalyzerConfigCreate(stopword_preset="custom"),
+            stopword_presets={"custom": ["test"]},
         )
         assert result.indexed == ["hello", "world", "test"]
         assert result.query == ["hello", "world"]
 
-    def test_stopword_presets_with_base_and_removals(self, client: weaviate.WeaviateClient) -> None:
-        cfg = _TextAnalyzerConfigCreate(stopword_preset="en-no-the")
+    def test_stopword_presets_override_builtin_en(
+        self, client: weaviate.WeaviateClient
+    ) -> None:
+        """A user-defined preset sharing a name with a built-in replaces the
+        built-in entirely, including on the default-en path for word
+        tokenization."""
         result = client.tokenization.text(
-            text="the quick",
+            text="the quick hello world",
             tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
-            stopword_presets={
-                "en-no-the": _StopwordsCreate(
-                    preset=StopwordsPreset.EN, additions=None, removals=["the"]
-                ),
-            },
+            stopword_presets={"en": ["hello"]},
         )
-        assert result.indexed == ["the", "quick"]
-        assert result.query == ["the", "quick"]
+        assert result.indexed == ["the", "quick", "hello", "world"]
+        # "the" no longer filtered (built-in en replaced), "hello" is.
+        assert result.query == ["the", "quick", "world"]
 
 
 # ---------------------------------------------------------------------------
@@ -176,61 +256,23 @@ def test_stopword_presets_with_base_and_removals(self, client: weaviate.Weaviate
 
 @pytest.mark.usefixtures("require_1_37")
 class TestDeserialization:
-    """Verify the client correctly deserializes response fields into typed objects."""
+    """Verify the client correctly deserializes response fields into
+    TokenizeResult."""
 
-    def test_result_type(self, client: weaviate.WeaviateClient) -> None:
+    def test_generic_result_shape(self, client: weaviate.WeaviateClient) -> None:
+        """Generic endpoint returns only indexed and query; tokenization is
+        not echoed back."""
         result = client.tokenization.text(text="hello", tokenization=Tokenization.WORD)
         assert isinstance(result, TokenizeResult)
         assert isinstance(result.indexed, list)
         assert isinstance(result.query, list)
+        assert result.tokenization is None
 
-    def test_analyzer_config_deserialized(self, client: weaviate.WeaviateClient) -> None:
-        cfg = _TextAnalyzerConfigCreate(
-            ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset=StopwordsPreset.EN
-        )
-        result = client.tokenization.text(
-            text="L'école",
-            tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
-        )
-        assert isinstance(result.analyzer_config, TextAnalyzerConfig)
-        assert result.analyzer_config.ascii_fold is True
-        assert result.analyzer_config.ascii_fold_ignore == ["é"]
-        assert result.analyzer_config.stopword_preset == "en"
-
-    def test_no_analyzer_config_returns_none(self, client: weaviate.WeaviateClient) -> None:
-        result = client.tokenization.text(text="hello", tokenization=Tokenization.WORD)
-        assert result.analyzer_config is None
-
-    def test_stopword_config_deserialized_on_property(
+    def test_property_result_populates_tokenization(
         self, client: weaviate.WeaviateClient
     ) -> None:
-        client.collections.delete("TestDeserStopword")
-        try:
-            client.collections.create_from_dict(
-                {
-                    "class": "TestDeserStopword",
-                    "vectorizer": "none",
-                    "properties": [
-                        {
-                            "name": "title",
-                            "dataType": ["text"],
-                            "tokenization": "word",
-                            "textAnalyzer": {"stopwordPreset": "en"},
-                        },
-                    ],
-                }
-            )
-            col = client.collections.get("TestDeserStopword")
-            result = col.config.tokenize_property(property_name="title", text="the quick")
-            assert isinstance(result, TokenizeResult)
-            assert result.tokenization == Tokenization.WORD
-            if result.stopword_config is not None:
-                assert isinstance(result.stopword_config, StopwordsConfig)
-        finally:
-            client.collections.delete("TestDeserStopword")
-
-    def test_property_result_types(self, client: weaviate.WeaviateClient) -> None:
+        """Property endpoint returns tokenization — the server resolved it
+        from the property's schema rather than the caller sending it."""
         client.collections.delete("TestDeserPropTypes")
         try:
             client.collections.create_from_dict(
@@ -256,12 +298,13 @@ def test_property_result_types(self, client: weaviate.WeaviateClient) -> None:
 
 
 # ---------------------------------------------------------------------------
-# Client-side validation (_TextAnalyzerConfigCreate)
+# Client-side validation
 # ---------------------------------------------------------------------------
 
 
 class TestClientSideValidation:
-    """Verify that _TextAnalyzerConfigCreate rejects invalid input before hitting the server."""
+    """Verify that client-side validation rejects invalid input before
+    hitting the server."""
 
     def test_ascii_fold_ignore_without_fold_raises(self) -> None:
         with pytest.raises(ValueError, match="asciiFoldIgnore"):
@@ -291,6 +334,23 @@ def test_empty_config_is_valid(self) -> None:
         assert cfg.asciiFoldIgnore is None
         assert cfg.stopwordPreset is None
 
+    def test_stopwords_and_stopword_presets_mutex(
+        self, client: weaviate.WeaviateClient
+    ) -> None:
+        """Client rejects the mutex violation locally with ValueError, before
+        sending the request (which the server would also reject with 422)."""
+        if client._connection._weaviate_version.is_lower_than(1, 37, 0):
+            pytest.skip("Tokenization requires Weaviate >= 1.37.0")
+        with pytest.raises(ValueError, match="mutually exclusive"):
+            client.tokenization.text(
+                text="hello",
+                tokenization=Tokenization.WORD,
+                stopwords=_StopwordsCreate(
+                    preset=StopwordsPreset.EN, additions=None, removals=None
+                ),
+                stopword_presets={"custom": ["hello"]},
+            )
+
 
 # ---------------------------------------------------------------------------
 # Version gate
@@ -331,20 +391,21 @@ async def test_text_tokenize(self, async_client: weaviate.WeaviateAsyncClient) -
         )
         assert isinstance(result, TokenizeResult)
         assert result.indexed == ["the", "quick", "brown", "fox"]
+        # default "en" applied server-side.
+        assert result.query == ["quick", "brown", "fox"]
 
     @pytest.mark.asyncio
-    async def test_text_with_analyzer_config(
+    async def test_text_with_stopwords_fallback(
         self, async_client: weaviate.WeaviateAsyncClient
     ) -> None:
-        cfg = _TextAnalyzerConfigCreate(ascii_fold=True, stopword_preset=StopwordsPreset.EN)
+        sw = _StopwordsCreate(preset=StopwordsPreset.EN, additions=["quick"], removals=None)
         result = await async_client.tokenization.text(
-            text="L'école est fermée",
+            text="the quick brown fox",
             tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
+            stopwords=sw,
         )
-        assert result.indexed == ["l", "ecole", "est", "fermee"]
-        assert isinstance(result.analyzer_config, TextAnalyzerConfig)
-        assert result.analyzer_config.ascii_fold is True
+        assert result.indexed == ["the", "quick", "brown", "fox"]
+        assert result.query == ["brown", "fox"]
 
     @pytest.mark.asyncio
     async def test_property_tokenize(self, async_client: weaviate.WeaviateAsyncClient) -> None:
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index 5093c14e9..3a79d6ee1 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -1,6 +1,6 @@
 """Tokenize executor."""
 
-from typing import Any, Dict, Generic, Optional
+from typing import Any, Dict, Generic, List, Optional
 
 from httpx import Response
 
@@ -33,26 +33,52 @@ def text(
         tokenization: Tokenization,
         *,
         analyzer_config: Optional[_TextAnalyzerConfigCreate] = None,
-        stopword_presets: Optional[Dict[str, _StopwordsCreate]] = None,
+        stopwords: Optional[_StopwordsCreate] = None,
+        stopword_presets: Optional[Dict[str, List[str]]] = None,
     ) -> executor.Result[TokenizeResult]:
         """Tokenize text using the generic /v1/tokenize endpoint.
 
+        For ``word`` tokenization the server defaults to the built-in ``en``
+        stopword preset when no stopword configuration is supplied. Pass
+        ``analyzer_config=TextAnalyzerConfig(stopword_preset="none")`` or
+        equivalent to opt out.
+
         Args:
             text: The text to tokenize.
             tokenization: The tokenization method to use (e.g. Tokenization.WORD).
-            analyzer_config: Text analyzer settings (ASCII folding, stopword preset).
-            stopword_presets: Custom stopword preset definitions, keyed by name.
-                Each value is a ``_StopwordsCreate`` with optional preset, additions,
-                and removals fields.
+            analyzer_config: Text analyzer settings (ASCII folding, stopword
+                preset name). ``stopword_preset`` may reference a built-in preset
+                (``en`` / ``none``) or a name defined in ``stopword_presets``.
+            stopwords: Fallback stopword config applied when
+                ``analyzer_config.stopword_preset`` is not set. Same shape as a
+                collection's ``invertedIndexConfig.stopwords`` — a base preset
+                optionally tweaked with ``additions`` / ``removals``. An empty
+                ``preset`` defaults to ``en``.
+            stopword_presets: User-defined named stopword presets, each a plain
+                list of words. A name matching a built-in (``en`` / ``none``)
+                replaces the built-in entirely.
+
+        Note:
+            ``stopwords`` and ``stopword_presets`` are mutually exclusive on the
+            server — pass one or the other, not both. The server returns HTTP
+            422 if both are supplied.
 
         Returns:
-            A TokenizeResult with indexed and query token lists.
+            A TokenizeResult with indexed and query token lists. The generic
+            endpoint does not echo request fields (tokenization, analyzer_config,
+            stopwords, stopword_presets) back in the response.
 
         Raises:
             WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
+            ValueError: If both ``stopwords`` and ``stopword_presets`` are passed.
         """
         self._check_version()
 
+        if stopwords is not None and stopword_presets is not None:
+            raise ValueError(
+                "stopwords and stopword_presets are mutually exclusive; pass only one"
+            )
+
         payload: Dict[str, Any] = {
             "text": text,
             "tokenization": tokenization.value,
@@ -63,9 +89,16 @@ def text(
             if ac_dict:
                 payload["analyzerConfig"] = ac_dict
 
+        if stopwords is not None:
+            sw_dict = stopwords._to_dict()
+            if sw_dict:
+                payload["stopwords"] = sw_dict
+
         if stopword_presets is not None:
+            # Plain word-list shape matching a collection's
+            # invertedIndexConfig.stopwordPresets.
             payload["stopwordPresets"] = {
-                name: cfg._to_dict() for name, cfg in stopword_presets.items()
+                name: list(words) for name, words in stopword_presets.items()
             }
 
         def resp(response: Response) -> TokenizeResult:
diff --git a/weaviate/tokenization/models.py b/weaviate/tokenization/models.py
index 8bfa508f8..3bf980597 100644
--- a/weaviate/tokenization/models.py
+++ b/weaviate/tokenization/models.py
@@ -1,56 +1,26 @@
 """Return types for tokenization operations."""
 
-from typing import Any, Dict, List, Optional
+from typing import List, Optional
 
-from pydantic import BaseModel, ConfigDict, Field, field_validator
+from pydantic import BaseModel, ConfigDict
 
-from weaviate.collections.classes.config import (
-    StopwordsConfig,
-    StopwordsPreset,
-    TextAnalyzerConfig,
-    Tokenization,
-)
+from weaviate.collections.classes.config import Tokenization
 
 
 class TokenizeResult(BaseModel):
     """Result of a tokenization operation.
 
     Attributes:
-        tokenization: The tokenization method that was applied.
         indexed: Tokens as they would be stored in the inverted index.
         query: Tokens as they would be used for querying (after stopword removal).
-        analyzer_config: The text analyzer configuration that was used, if any.
-        stopword_config: The stopword configuration that was used, if any.
+        tokenization: The tokenization method that was applied. Populated only by
+            the property-level endpoint, where the tokenization is resolved from
+            the property's schema. The generic ``/v1/tokenize`` endpoint does not
+            echo it back (the caller passed it).
     """
 
     model_config = ConfigDict(populate_by_name=True, arbitrary_types_allowed=True)
 
-    tokenization: Tokenization
     indexed: List[str]
     query: List[str]
-    analyzer_config: Optional[TextAnalyzerConfig] = Field(default=None, alias="analyzerConfig")
-    stopword_config: Optional[StopwordsConfig] = Field(default=None, alias="stopwordConfig")
-
-    @field_validator("analyzer_config", mode="before")
-    @classmethod
-    def _parse_analyzer_config(cls, v: Optional[Dict[str, Any]]) -> Optional[TextAnalyzerConfig]:
-        if v is None:
-            return None
-        if "asciiFold" not in v and "stopwordPreset" not in v:
-            return None
-        return TextAnalyzerConfig(
-            ascii_fold=v.get("asciiFold", False),
-            ascii_fold_ignore=v.get("asciiFoldIgnore"),
-            stopword_preset=v.get("stopwordPreset"),
-        )
-
-    @field_validator("stopword_config", mode="before")
-    @classmethod
-    def _parse_stopword_config(cls, v: Optional[Dict[str, Any]]) -> Optional[StopwordsConfig]:
-        if v is None:
-            return None
-        return StopwordsConfig(
-            preset=StopwordsPreset(v["preset"]),
-            additions=v.get("additions"),
-            removals=v.get("removals"),
-        )
+    tokenization: Optional[Tokenization] = None

From 5a12f134c36a5f6ed6329abaae101abf42efb7cf Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 20 Apr 2026 15:55:45 +0100
Subject: [PATCH 85/99] fix: update Weaviate 1.37.1 version to include specific
 build identifier

---
 .github/workflows/main.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index 8dd157443..ee9b69537 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -28,7 +28,7 @@ env:
   WEAVIATE_134: 1.34.19
   WEAVIATE_135: 1.35.16-efdedfa
   WEAVIATE_136: 1.36.9-d905e6c
-  WEAVIATE_137: 1.37.1
+  WEAVIATE_137: 1.37.1-5f911bc
 
 jobs:
   lint-and-format:

From d7605777b43a799eb31ff712e642a97ee63db855 Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Mon, 20 Apr 2026 18:07:58 +0100
Subject: [PATCH 86/99] Reduce timeouts in batch tests

---
 integration/test_batch_v4.py         | 6 +++---
 integration/test_collection_batch.py | 6 +++---
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/integration/test_batch_v4.py b/integration/test_batch_v4.py
index f4ce7669e..3646f1b1f 100644
--- a/integration/test_batch_v4.py
+++ b/integration/test_batch_v4.py
@@ -718,7 +718,7 @@ def test_non_existant_collection(client_factory: ClientFactory) -> None:
     # not, so we do not check for errors here
 
 
-@pytest.mark.timeout(600)
+@pytest.mark.timeout(60)
 def test_number_of_stored_results_in_batch(client_factory: ClientFactory) -> None:
     client, name = client_factory()
     with client.batch.dynamic() as batch:
@@ -818,7 +818,7 @@ def test_references_with_to_uuids(client_factory: ClientFactory) -> None:
 
 
 @pytest.mark.asyncio
-@pytest.mark.timeout(600)
+@pytest.mark.timeout(60)
 async def test_add_one_hundred_thousand_objects_async_client(
     async_client_factory: AsyncClientFactory,
 ) -> None:
@@ -849,7 +849,7 @@ async def test_add_one_hundred_thousand_objects_async_client(
     await client.collections.delete(name)
 
 
-@pytest.mark.timeout(600)
+@pytest.mark.timeout(60)
 def test_add_one_hundred_thousand_objects_sync_client(
     client_factory: ClientFactory,
 ) -> None:
diff --git a/integration/test_collection_batch.py b/integration/test_collection_batch.py
index e670e4883..f2bd5be61 100644
--- a/integration/test_collection_batch.py
+++ b/integration/test_collection_batch.py
@@ -271,7 +271,7 @@ def test_non_existant_collection(collection_factory_get: CollectionFactoryGet) -
 
 
 @pytest.mark.asyncio
-@pytest.mark.timeout(600)
+@pytest.mark.timeout(60)
 async def test_batch_one_hundred_thousand_objects_async_collection(
     batch_collection_async: BatchCollectionAsync,
 ) -> None:
@@ -299,7 +299,7 @@ async def test_batch_one_hundred_thousand_objects_async_collection(
 
 
 @pytest.mark.asyncio
-@pytest.mark.timeout(600)
+@pytest.mark.timeout(60)
 async def test_ingest_one_hundred_thousand_data_objects_async(
     batch_collection_async: BatchCollectionAsync,
 ) -> None:
@@ -321,7 +321,7 @@ async def test_ingest_one_hundred_thousand_data_objects_async(
     assert len(results.errors) == 0, [obj.message for obj in results.errors.values()]
 
 
-@pytest.mark.timeout(600)
+@pytest.mark.timeout(60)
 def test_ingest_one_hundred_thousand_data_objects(
     batch_collection: BatchCollection,
 ) -> None:

From 60887f3ab37171d2ea12aa34de57a927fcca8267 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Mon, 20 Apr 2026 18:08:38 +0100
Subject: [PATCH 87/99] fix: update Weaviate 1.37.1 version to include
 architecture suffix

---
 .github/workflows/main.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index b2c567cee..94f75b089 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -28,7 +28,7 @@ env:
   WEAVIATE_134: 1.34.19
   WEAVIATE_135: 1.35.18
   WEAVIATE_136: 1.36.12
-  WEAVIATE_137: 1.37.1-5f911bc
+  WEAVIATE_137: 1.37.1-5f911bc.amd64
 
 jobs:
   lint-and-format:

From 9fd83b881c953f2ea09c1fb5c89dc38a96173c29 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 21 Apr 2026 09:23:37 +0100
Subject: [PATCH 88/99] fix: refactor tokenization tests to use parameterized
 cases for improved readability and maintainability

---
 integration/test_tokenize.py | 327 +++++++++++++++++++----------------
 1 file changed, 176 insertions(+), 151 deletions(-)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index 788cefc31..51f154479 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -109,131 +109,146 @@ def test_tokenization_enum(
         # Generic endpoint does not echo tokenization back.
         assert result.tokenization is None
 
-    def test_default_en_applied_for_word(self, client: weaviate.WeaviateClient) -> None:
-        """Word tokenization defaults to the 'en' preset when no stopword config is supplied."""
-        result = client.tokenization.text(
-            text="The quick brown fox", tokenization=Tokenization.WORD
-        )
-        assert result.indexed == ["the", "quick", "brown", "fox"]
-        # "the" removed by the server's default en preset.
-        assert result.query == ["quick", "brown", "fox"]
-
-    def test_opt_out_of_default_en(self, client: weaviate.WeaviateClient) -> None:
-        """analyzerConfig.stopwordPreset='none' disables the default en."""
-        cfg = _TextAnalyzerConfigCreate(stopword_preset=StopwordsPreset.NONE)
-        result = client.tokenization.text(
-            text="The quick brown fox",
-            tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
-        )
-        assert result.query == ["the", "quick", "brown", "fox"]
-
-    def test_ascii_fold(self, client: weaviate.WeaviateClient) -> None:
-        cfg = _TextAnalyzerConfigCreate(ascii_fold=True)
-        result = client.tokenization.text(
-            text="L'école est fermée",
-            tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
-        )
-        assert result.indexed == ["l", "ecole", "est", "fermee"]
-
-    def test_ascii_fold_with_ignore(self, client: weaviate.WeaviateClient) -> None:
-        cfg = _TextAnalyzerConfigCreate(ascii_fold=True, ascii_fold_ignore=["é"])
-        result = client.tokenization.text(
-            text="L'école est fermée",
-            tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
-        )
-        assert result.indexed == ["l", "école", "est", "fermée"]
-
-    def test_stopword_preset_enum(self, client: weaviate.WeaviateClient) -> None:
-        cfg = _TextAnalyzerConfigCreate(stopword_preset=StopwordsPreset.EN)
-        result = client.tokenization.text(
-            text="The quick brown fox",
-            tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
-        )
-        assert "the" not in result.query
-        assert "quick" in result.query
-
-    def test_stopword_preset_string(self, client: weaviate.WeaviateClient) -> None:
-        cfg = _TextAnalyzerConfigCreate(stopword_preset="en")
-        result = client.tokenization.text(
-            text="The quick brown fox",
-            tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
-        )
-        assert "the" not in result.query
-
-    def test_ascii_fold_combined_with_stopwords(self, client: weaviate.WeaviateClient) -> None:
-        cfg = _TextAnalyzerConfigCreate(
-            ascii_fold=True, ascii_fold_ignore=["é"], stopword_preset=StopwordsPreset.EN
-        )
-        result = client.tokenization.text(
-            text="The école est fermée",
-            tokenization=Tokenization.WORD,
-            analyzer_config=cfg,
-        )
-        assert result.indexed == ["the", "école", "est", "fermée"]
-        assert "the" not in result.query
-        assert "école" in result.query
-
-    def test_stopwords_fallback(self, client: weaviate.WeaviateClient) -> None:
-        """Top-level stopwords acts as the fallback detector when no analyzerConfig.stopwordPreset is set."""
-        sw = _StopwordsCreate(preset=StopwordsPreset.EN, additions=["quick"], removals=None)
-        result = client.tokenization.text(
-            text="the quick brown fox",
-            tokenization=Tokenization.WORD,
-            stopwords=sw,
-        )
-        assert result.indexed == ["the", "quick", "brown", "fox"]
-        # "the" (en) and "quick" (addition) filtered.
-        assert result.query == ["brown", "fox"]
-
-    def test_stopwords_additions_default_preset_to_en(
-        self, client: weaviate.WeaviateClient
+    @pytest.mark.parametrize(
+        "call_kwargs,expected_indexed,expected_query",
+        [
+            (
+                {"text": "The quick brown fox"},
+                ["the", "quick", "brown", "fox"],
+                ["quick", "brown", "fox"],
+            ),
+            (
+                {
+                    "text": "The quick brown fox",
+                    "analyzer_config": _TextAnalyzerConfigCreate(
+                        stopword_preset=StopwordsPreset.NONE
+                    ),
+                },
+                ["the", "quick", "brown", "fox"],
+                ["the", "quick", "brown", "fox"],
+            ),
+            (
+                {
+                    "text": "L'école est fermée",
+                    "analyzer_config": _TextAnalyzerConfigCreate(ascii_fold=True),
+                },
+                ["l", "ecole", "est", "fermee"],
+                ["l", "ecole", "fermee"],
+            ),
+            (
+                {
+                    "text": "L'école est fermée",
+                    "analyzer_config": _TextAnalyzerConfigCreate(
+                        ascii_fold=True, ascii_fold_ignore=["é"]
+                    ),
+                },
+                ["l", "école", "est", "fermée"],
+                ["l", "école", "fermée"],
+            ),
+            (
+                {
+                    "text": "The quick brown fox",
+                    "analyzer_config": _TextAnalyzerConfigCreate(
+                        stopword_preset=StopwordsPreset.EN
+                    ),
+                },
+                ["the", "quick", "brown", "fox"],
+                ["quick", "brown", "fox"],
+            ),
+            (
+                {
+                    "text": "The quick brown fox",
+                    "analyzer_config": _TextAnalyzerConfigCreate(stopword_preset="en"),
+                },
+                ["the", "quick", "brown", "fox"],
+                ["quick", "brown", "fox"],
+            ),
+            (
+                {
+                    "text": "The école est fermée",
+                    "analyzer_config": _TextAnalyzerConfigCreate(
+                        ascii_fold=True,
+                        ascii_fold_ignore=["é"],
+                        stopword_preset=StopwordsPreset.EN,
+                    ),
+                },
+                ["the", "école", "est", "fermée"],
+                ["école", "est", "fermée"],
+            ),
+            (
+                {
+                    "text": "the quick brown fox",
+                    "stopwords": _StopwordsCreate(
+                        preset=StopwordsPreset.EN, additions=["quick"], removals=None
+                    ),
+                },
+                ["the", "quick", "brown", "fox"],
+                ["brown", "fox"],
+            ),
+            (
+                {
+                    "text": "the quick hello world",
+                    "stopwords": _StopwordsCreate(
+                        preset=None, additions=["hello"], removals=None
+                    ),
+                },
+                ["the", "quick", "hello", "world"],
+                ["quick", "world"],
+            ),
+            (
+                {
+                    "text": "the quick is fast",
+                    "stopwords": _StopwordsCreate(
+                        preset=None, additions=None, removals=["the"]
+                    ),
+                },
+                ["the", "quick", "is", "fast"],
+                ["the", "quick", "fast"],
+            ),
+            (
+                {
+                    "text": "hello world test",
+                    "analyzer_config": _TextAnalyzerConfigCreate(stopword_preset="custom"),
+                    "stopword_presets": {"custom": ["test"]},
+                },
+                ["hello", "world", "test"],
+                ["hello", "world"],
+            ),
+            (
+                {
+                    "text": "the quick hello world",
+                    "stopword_presets": {"en": ["hello"]},
+                },
+                ["the", "quick", "hello", "world"],
+                ["the", "quick", "world"],
+            ),
+        ],
+        ids=[
+            "default_en_applied_for_word",
+            "opt_out_of_default_en",
+            "ascii_fold",
+            "ascii_fold_with_ignore",
+            "stopword_preset_enum",
+            "stopword_preset_string",
+            "ascii_fold_combined_with_stopwords",
+            "stopwords_fallback",
+            "stopwords_additions_default_preset_to_en",
+            "stopwords_removals_default_preset_to_en",
+            "stopword_presets_named_reference",
+            "stopword_presets_override_builtin_en",
+        ],
+    )
+    def test_text_tokenize(
+        self,
+        client: weaviate.WeaviateClient,
+        call_kwargs: dict,
+        expected_indexed: list,
+        expected_query: list,
     ) -> None:
-        """Caller omits preset, passes only additions. Server defaults preset to 'en' and builds detector from en + additions."""
-        sw = _StopwordsCreate(preset=None, additions=["hello"], removals=None)
-        result = client.tokenization.text(
-            text="the quick hello world",
-            tokenization=Tokenization.WORD,
-            stopwords=sw,
-        )
-        assert result.query == ["quick", "world"]
-
-    def test_stopwords_removals_default_preset_to_en(self, client: weaviate.WeaviateClient) -> None:
-        """Caller omits preset, passes only removals. 'the' is removed from the en list so it passes through."""
-        sw = _StopwordsCreate(preset=None, additions=None, removals=["the"])
-        result = client.tokenization.text(
-            text="the quick is fast",
-            tokenization=Tokenization.WORD,
-            stopwords=sw,
-        )
-        # "is" still in en, "the" removed.
-        assert result.query == ["the", "quick", "fast"]
-
-    def test_stopword_presets_named_reference(self, client: weaviate.WeaviateClient) -> None:
-        """Define a named preset via stopword_presets, select it via analyzerConfig.stopwordPreset. Word lists use the collection shape."""
-        result = client.tokenization.text(
-            text="hello world test",
-            tokenization=Tokenization.WORD,
-            analyzer_config=_TextAnalyzerConfigCreate(stopword_preset="custom"),
-            stopword_presets={"custom": ["test"]},
-        )
-        assert result.indexed == ["hello", "world", "test"]
-        assert result.query == ["hello", "world"]
-
-    def test_stopword_presets_override_builtin_en(self, client: weaviate.WeaviateClient) -> None:
-        """A user-defined preset sharing a name with a built-in replaces the built-in entirely, including on the default-en path for word tokenization."""
-        result = client.tokenization.text(
-            text="the quick hello world",
-            tokenization=Tokenization.WORD,
-            stopword_presets={"en": ["hello"]},
-        )
-        assert result.indexed == ["the", "quick", "hello", "world"]
-        # "the" no longer filtered (built-in en replaced), "hello" is.
-        assert result.query == ["the", "quick", "world"]
+        result = client.tokenization.text(tokenization=Tokenization.WORD, **call_kwargs)
+        assert isinstance(result, TokenizeResult)
+        assert result.indexed == expected_indexed
+        assert result.query == expected_query
 
 
 # ---------------------------------------------------------------------------
@@ -287,33 +302,44 @@ def test_property_result_populates_tokenization(self, client: weaviate.WeaviateC
 class TestClientSideValidation:
     """Verify that client-side validation rejects invalid input before hitting the server."""
 
-    def test_ascii_fold_ignore_without_fold_raises(self) -> None:
-        with pytest.raises(ValueError, match="asciiFoldIgnore"):
-            _TextAnalyzerConfigCreate(ascii_fold=False, ascii_fold_ignore=["é"])
-
-    def test_ascii_fold_ignore_without_fold_default_raises(self) -> None:
+    @pytest.mark.parametrize(
+        "kwargs",
+        [
+            {"ascii_fold": False, "ascii_fold_ignore": ["é"]},
+            {"ascii_fold_ignore": ["é"]},
+        ],
+        ids=["explicit_false", "default"],
+    )
+    def test_ascii_fold_ignore_without_fold_raises(self, kwargs: dict) -> None:
         with pytest.raises(ValueError, match="asciiFoldIgnore"):
-            _TextAnalyzerConfigCreate(ascii_fold_ignore=["é"])
+            _TextAnalyzerConfigCreate(**kwargs)
 
-    def test_valid_config_does_not_raise(self) -> None:
-        cfg = _TextAnalyzerConfigCreate(ascii_fold=True, ascii_fold_ignore=["é", "ñ"])
-        assert cfg.asciiFold is True
-        assert cfg.asciiFoldIgnore == ["é", "ñ"]
-
-    def test_fold_without_ignore_is_valid(self) -> None:
-        cfg = _TextAnalyzerConfigCreate(ascii_fold=True)
-        assert cfg.asciiFold is True
-        assert cfg.asciiFoldIgnore is None
-
-    def test_stopword_preset_only_is_valid(self) -> None:
-        cfg = _TextAnalyzerConfigCreate(stopword_preset="en")
-        assert cfg.stopwordPreset == "en"
-
-    def test_empty_config_is_valid(self) -> None:
-        cfg = _TextAnalyzerConfigCreate()
-        assert cfg.asciiFold is None
-        assert cfg.asciiFoldIgnore is None
-        assert cfg.stopwordPreset is None
+    @pytest.mark.parametrize(
+        "kwargs,expected",
+        [
+            (
+                {"ascii_fold": True, "ascii_fold_ignore": ["é", "ñ"]},
+                {"asciiFold": True, "asciiFoldIgnore": ["é", "ñ"]},
+            ),
+            (
+                {"ascii_fold": True},
+                {"asciiFold": True, "asciiFoldIgnore": None},
+            ),
+            (
+                {"stopword_preset": "en"},
+                {"stopwordPreset": "en"},
+            ),
+            (
+                {},
+                {"asciiFold": None, "asciiFoldIgnore": None, "stopwordPreset": None},
+            ),
+        ],
+        ids=["fold_with_ignore", "fold_without_ignore", "stopword_preset_only", "empty"],
+    )
+    def test_valid_config(self, kwargs: dict, expected: dict) -> None:
+        cfg = _TextAnalyzerConfigCreate(**kwargs)
+        for attr, value in expected.items():
+            assert getattr(cfg, attr) == value
 
     def test_stopwords_and_stopword_presets_mutex(self, client: weaviate.WeaviateClient) -> None:
         """Client rejects the mutex violation locally with ValueError, before sending the request (which the server would also reject with 422)."""
@@ -411,7 +437,6 @@ async def test_property_tokenize(self, async_client: weaviate.WeaviateAsyncClien
             assert isinstance(result, TokenizeResult)
             assert result.tokenization == Tokenization.WORD
             assert result.indexed == ["the", "quick", "brown", "fox"]
-            assert "the" not in result.query
-            assert "quick" in result.query
+            assert result.query == ["quick", "brown", "fox"]
         finally:
             await async_client.collections.delete("TestAsyncPropTokenize")

From e9d681226e1917b80fed0312b086a5818cac2e9c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 21 Apr 2026 09:52:24 +0100
Subject: [PATCH 89/99] fix: update Weaviate 1.37.1 version and enhance
 tokenization tests with new fixtures

---
 .github/workflows/main.yaml  |  2 +-
 integration/test_tokenize.py | 77 ++++++++++++++++++++++++++++++------
 2 files changed, 66 insertions(+), 13 deletions(-)

diff --git a/.github/workflows/main.yaml b/.github/workflows/main.yaml
index 94f75b089..a1ff94f98 100644
--- a/.github/workflows/main.yaml
+++ b/.github/workflows/main.yaml
@@ -28,7 +28,7 @@ env:
   WEAVIATE_134: 1.34.19
   WEAVIATE_135: 1.35.18
   WEAVIATE_136: 1.36.12
-  WEAVIATE_137: 1.37.1-5f911bc.amd64
+  WEAVIATE_137: 1.37.1-4e61e26.amd64
 
 jobs:
   lint-and-format:
diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index 51f154479..61d54e095 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -9,8 +9,7 @@
 
 Server-side behavior this client relies on:
 - Word tokenization defaults to preset "en" when no stopword config is sent.
-- The generic /v1/tokenize response is minimal: only ``indexed`` and ``query``
-  are returned. The property-level endpoint additionally returns ``tokenization``.
+- Both endpoints return only ``indexed`` and ``query``.
 - ``stopwords`` and ``stopword_presets`` are mutually exclusive on the generic
   endpoint — the server rejects requests that set both.
 """
@@ -57,6 +56,29 @@ async def async_client() -> AsyncGenerator[weaviate.WeaviateAsyncClient, None]:
     await c.close()
 
 
+@pytest.fixture
+def recipe_collection(client: weaviate.WeaviateClient) -> Generator:
+    """Collection with a `recipe` word-tokenized property and an en + ["quick"] stopwords config."""
+    name = "TestTokenizeRecipe"
+    client.collections.delete(name)
+    client.collections.create_from_dict(
+        {
+            "class": name,
+            "vectorizer": "none",
+            "invertedIndexConfig": {
+                "stopwords": {"preset": "en", "additions": ["quick"]},
+            },
+            "properties": [
+                {"name": "recipe", "dataType": ["text"], "tokenization": "word"},
+            ],
+        }
+    )
+    try:
+        yield client.collections.get(name)
+    finally:
+        client.collections.delete(name)
+
+
 # ---------------------------------------------------------------------------
 # Serialization
 # ---------------------------------------------------------------------------
@@ -106,8 +128,6 @@ def test_tokenization_enum(
         assert isinstance(result, TokenizeResult)
         assert result.indexed == expected_indexed
         assert result.query == expected_query
-        # Generic endpoint does not echo tokenization back.
-        assert result.tokenization is None
 
     @pytest.mark.parametrize(
         "call_kwargs,expected_indexed,expected_query",
@@ -133,7 +153,7 @@ def test_tokenization_enum(
                     "analyzer_config": _TextAnalyzerConfigCreate(ascii_fold=True),
                 },
                 ["l", "ecole", "est", "fermee"],
-                ["l", "ecole", "fermee"],
+                ["l", "ecole", "est", "fermee"],
             ),
             (
                 {
@@ -143,7 +163,7 @@ def test_tokenization_enum(
                     ),
                 },
                 ["l", "école", "est", "fermée"],
-                ["l", "école", "fermée"],
+                ["l", "école", "est", "fermée"],
             ),
             (
                 {
@@ -250,6 +270,42 @@ def test_text_tokenize(
         assert result.indexed == expected_indexed
         assert result.query == expected_query
 
+    def test_text_from_collection_config(
+        self, client: weaviate.WeaviateClient, recipe_collection
+    ) -> None:
+        """Values round-tripped through config.get() feed back into tokenization.text()."""
+        config = recipe_collection.config.get()
+        recipe = next(p for p in config.properties if p.name == "recipe")
+        stopwords = config.inverted_index_config.stopwords
+        result = client.tokenization.text(
+            text="the quick brown fox",
+            tokenization=recipe.tokenization,
+            stopwords=_StopwordsCreate(**stopwords.__dict__),
+        )
+        assert result.indexed == ["the", "quick", "brown", "fox"]
+        assert result.query == ["brown", "fox"]
+
+    def test_property_and_generic_endpoints_agree(
+        self, client: weaviate.WeaviateClient, recipe_collection
+    ) -> None:
+        """Property endpoint (server resolves config from schema) produces the same indexed/query as the generic endpoint fed the same config."""
+        config = recipe_collection.config.get()
+        recipe = next(p for p in config.properties if p.name == "recipe")
+        stopwords = config.inverted_index_config.stopwords
+
+        text = "the quick brown fox"
+        via_property = recipe_collection.config.tokenize_property(
+            property_name="recipe", text=text
+        )
+        via_generic = client.tokenization.text(
+            text=text,
+            tokenization=recipe.tokenization,
+            stopwords=_StopwordsCreate(**stopwords.__dict__),
+        )
+
+        assert via_property.indexed == via_generic.indexed
+        assert via_property.query == via_generic.query
+
 
 # ---------------------------------------------------------------------------
 # Deserialization
@@ -261,15 +317,14 @@ class TestDeserialization:
     """Verify the client correctly deserializes response fields into TokenizeResult."""
 
     def test_generic_result_shape(self, client: weaviate.WeaviateClient) -> None:
-        """Generic endpoint returns only indexed and query; tokenization is not echoed back."""
+        """Generic endpoint response deserializes into TokenizeResult with indexed and query lists."""
         result = client.tokenization.text(text="hello", tokenization=Tokenization.WORD)
         assert isinstance(result, TokenizeResult)
         assert isinstance(result.indexed, list)
         assert isinstance(result.query, list)
-        assert result.tokenization is None
 
-    def test_property_result_populates_tokenization(self, client: weaviate.WeaviateClient) -> None:
-        """Property endpoint returns tokenization — the server resolved it from the property's schema rather than the caller sending it."""
+    def test_property_result_shape(self, client: weaviate.WeaviateClient) -> None:
+        """Property endpoint response deserializes into TokenizeResult — server resolves tokenization from the property's schema."""
         client.collections.delete("TestDeserPropTypes")
         try:
             client.collections.create_from_dict(
@@ -288,7 +343,6 @@ def test_property_result_populates_tokenization(self, client: weaviate.WeaviateC
             col = client.collections.get("TestDeserPropTypes")
             result = col.config.tokenize_property(property_name="tag", text="  Hello World  ")
             assert isinstance(result, TokenizeResult)
-            assert result.tokenization == Tokenization.FIELD
             assert result.indexed == ["Hello World"]
         finally:
             client.collections.delete("TestDeserPropTypes")
@@ -435,7 +489,6 @@ async def test_property_tokenize(self, async_client: weaviate.WeaviateAsyncClien
                 text="The quick brown fox",
             )
             assert isinstance(result, TokenizeResult)
-            assert result.tokenization == Tokenization.WORD
             assert result.indexed == ["the", "quick", "brown", "fox"]
             assert result.query == ["quick", "brown", "fox"]
         finally:

From 959f554c7df129a226ddbc5f412e95a3879891d2 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 21 Apr 2026 09:55:35 +0100
Subject: [PATCH 90/99] refactor: ruff format

---
 integration/test_tokenize.py | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index 61d54e095..d2a8442d8 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -208,9 +208,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "the quick hello world",
-                    "stopwords": _StopwordsCreate(
-                        preset=None, additions=["hello"], removals=None
-                    ),
+                    "stopwords": _StopwordsCreate(preset=None, additions=["hello"], removals=None),
                 },
                 ["the", "quick", "hello", "world"],
                 ["quick", "world"],
@@ -218,9 +216,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "the quick is fast",
-                    "stopwords": _StopwordsCreate(
-                        preset=None, additions=None, removals=["the"]
-                    ),
+                    "stopwords": _StopwordsCreate(preset=None, additions=None, removals=["the"]),
                 },
                 ["the", "quick", "is", "fast"],
                 ["the", "quick", "fast"],
@@ -294,9 +290,7 @@ def test_property_and_generic_endpoints_agree(
         stopwords = config.inverted_index_config.stopwords
 
         text = "the quick brown fox"
-        via_property = recipe_collection.config.tokenize_property(
-            property_name="recipe", text=text
-        )
+        via_property = recipe_collection.config.tokenize_property(property_name="recipe", text=text)
         via_generic = client.tokenization.text(
             text=text,
             tokenization=recipe.tokenization,

From 0f7fe47cac92107f502fb5a22c925896084535ea Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 21 Apr 2026 14:24:33 +0100
Subject: [PATCH 91/99] test: refactor output types and tests to config

---
 integration/test_tokenize.py      | 32 ++++++++++++++++++++++++
 weaviate/tokenization/executor.py | 41 ++++++++++++++++++++++---------
 weaviate/tokenization/models.py   |  5 ----
 3 files changed, 62 insertions(+), 16 deletions(-)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index d2a8442d8..c939e8c5b 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -403,6 +403,38 @@ def test_stopwords_and_stopword_presets_mutex(self, client: weaviate.WeaviateCli
                 stopword_presets={"custom": ["hello"]},
             )
 
+    @pytest.mark.parametrize(
+        "stopword_presets,match",
+        [
+            ({"custom": "hello"}, "must be a list of strings"),
+            (
+                {
+                    "custom": _StopwordsCreate(
+                        preset=StopwordsPreset.EN, additions=None, removals=None
+                    ),
+                },
+                "must be a list of strings",
+            ),
+            ({"custom": ["hello", 123]}, "must contain only strings"),
+        ],
+        ids=["str_value", "pydantic_model_value", "non_string_element"],
+    )
+    def test_stopword_presets_invalid_shape_raises(
+        self,
+        client: weaviate.WeaviateClient,
+        stopword_presets: dict,
+        match: str,
+    ) -> None:
+        """Client rejects malformed stopword_presets values locally before sending — str would silently split into characters; a pydantic model would serialize to field tuples."""
+        if client._connection._weaviate_version.is_lower_than(1, 37, 0):
+            pytest.skip("Tokenization requires Weaviate >= 1.37.0")
+        with pytest.raises(ValueError, match=match):
+            client.tokenization.text(
+                text="hello",
+                tokenization=Tokenization.WORD,
+                stopword_presets=stopword_presets,
+            )
+
 
 # ---------------------------------------------------------------------------
 # Version gate
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index 8cedb6e51..825faee05 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -40,14 +40,15 @@ def text(
 
         For ``word`` tokenization the server defaults to the built-in ``en``
         stopword preset when no stopword configuration is supplied. Pass
-        ``analyzer_config=TextAnalyzerConfig(stopword_preset="none")`` or
-        equivalent to opt out.
+        ``analyzer_config=Configure.text_analyzer(stopword_preset=StopwordsPreset.NONE)``
+        (or equivalent) to opt out.
 
         Args:
             text: The text to tokenize.
             tokenization: The tokenization method to use (e.g. Tokenization.WORD).
             analyzer_config: Text analyzer settings (ASCII folding, stopword
-                preset name). ``stopword_preset`` may reference a built-in preset
+                preset name), built via ``Configure.text_analyzer(...)``.
+                ``stopword_preset`` may reference a built-in preset
                 (``en`` / ``none``) or a name defined in ``stopword_presets``.
             stopwords: Fallback stopword config applied when
                 ``analyzer_config.stopword_preset`` is not set. Same shape as a
@@ -64,13 +65,13 @@ def text(
             422 if both are supplied.
 
         Returns:
-            A TokenizeResult with indexed and query token lists. The generic
-            endpoint does not echo request fields (tokenization, analyzer_config,
-            stopwords, stopword_presets) back in the response.
+            A TokenizeResult with indexed and query token lists. The response
+            does not echo request fields back.
 
         Raises:
             WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
-            ValueError: If both ``stopwords`` and ``stopword_presets`` are passed.
+            ValueError: If both ``stopwords`` and ``stopword_presets`` are passed,
+                or if any ``stopword_presets`` value is not a list/tuple of strings.
         """
         self.__check_version()
 
@@ -94,10 +95,28 @@ def text(
 
         if stopword_presets is not None:
             # Plain word-list shape matching a collection's
-            # invertedIndexConfig.stopwordPresets.
-            payload["stopwordPresets"] = {
-                name: list(words) for name, words in stopword_presets.items()
-            }
+            # invertedIndexConfig.stopwordPresets. Reject str (would
+            # silently split into characters) and pydantic models /
+            # other non-sequence shapes up-front so callers get a clear
+            # error instead of a malformed payload.
+            validated: Dict[str, List[str]] = {}
+            for name, words in stopword_presets.items():
+                if isinstance(words, (str, bytes)):
+                    raise ValueError(
+                        f"stopword_presets[{name!r}] must be a list of strings, "
+                        f"got {type(words).__name__}"
+                    )
+                if not isinstance(words, (list, tuple)):
+                    raise ValueError(
+                        f"stopword_presets[{name!r}] must be a list of strings, "
+                        f"got {type(words).__name__}"
+                    )
+                if not all(isinstance(w, str) for w in words):
+                    raise ValueError(
+                        f"stopword_presets[{name!r}] must contain only strings"
+                    )
+                validated[name] = list(words)
+            payload["stopwordPresets"] = validated
 
         def resp(response: Response) -> TokenizeResult:
             return TokenizeResult.model_validate(response.json())
diff --git a/weaviate/tokenization/models.py b/weaviate/tokenization/models.py
index 3bf980597..017abe429 100644
--- a/weaviate/tokenization/models.py
+++ b/weaviate/tokenization/models.py
@@ -13,14 +13,9 @@ class TokenizeResult(BaseModel):
     Attributes:
         indexed: Tokens as they would be stored in the inverted index.
         query: Tokens as they would be used for querying (after stopword removal).
-        tokenization: The tokenization method that was applied. Populated only by
-            the property-level endpoint, where the tokenization is resolved from
-            the property's schema. The generic ``/v1/tokenize`` endpoint does not
-            echo it back (the caller passed it).
     """
 
     model_config = ConfigDict(populate_by_name=True, arbitrary_types_allowed=True)
 
     indexed: List[str]
     query: List[str]
-    tokenization: Optional[Tokenization] = None

From 52c2c8c8133eb1828be86f10ea824d74b822ca8a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Andr=C3=A9=20Mour=C3=A3o?= <andre.mourao@weaviate.io>
Date: Tue, 21 Apr 2026 14:31:17 +0100
Subject: [PATCH 92/99] refactor: remove unused imports in tokenization models
 and format

---
 weaviate/tokenization/executor.py | 4 +---
 weaviate/tokenization/models.py   | 4 +---
 2 files changed, 2 insertions(+), 6 deletions(-)

diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index 825faee05..25b36e1d3 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -112,9 +112,7 @@ def text(
                         f"got {type(words).__name__}"
                     )
                 if not all(isinstance(w, str) for w in words):
-                    raise ValueError(
-                        f"stopword_presets[{name!r}] must contain only strings"
-                    )
+                    raise ValueError(f"stopword_presets[{name!r}] must contain only strings")
                 validated[name] = list(words)
             payload["stopwordPresets"] = validated
 
diff --git a/weaviate/tokenization/models.py b/weaviate/tokenization/models.py
index 017abe429..f8fe7cb67 100644
--- a/weaviate/tokenization/models.py
+++ b/weaviate/tokenization/models.py
@@ -1,11 +1,9 @@
 """Return types for tokenization operations."""
 
-from typing import List, Optional
+from typing import List
 
 from pydantic import BaseModel, ConfigDict
 
-from weaviate.collections.classes.config import Tokenization
-
 
 class TokenizeResult(BaseModel):
     """Result of a tokenization operation.

From 3de0955c0520358b5d12f81b094b98ef3d208559 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 21 Apr 2026 15:07:26 +0200
Subject: [PATCH 93/99] Use public classes for .text endpoint

---
 integration/test_tokenize.py           | 48 ++++++++++++--------------
 weaviate/classes/config.py             |  4 +++
 weaviate/classes/tokenization.py       | 15 ++++++++
 weaviate/collections/classes/config.py |  4 +++
 weaviate/tokenization/async_.pyi       |  8 ++---
 weaviate/tokenization/executor.py      |  8 ++---
 weaviate/tokenization/sync.pyi         |  8 ++---
 7 files changed, 57 insertions(+), 38 deletions(-)
 create mode 100644 weaviate/classes/tokenization.py

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index c939e8c5b..a5b16da32 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -1,9 +1,9 @@
 """Integration tests for the tokenization module.
 
 These tests cover the client's responsibilities:
-- Correct serialization of inputs (enums, _TextAnalyzerConfigCreate, _StopwordsCreate)
+- Correct serialization of inputs (enums, TextAnalyzerConfigCreate, StopwordsCreate)
 - Correct deserialization of responses into the TokenizeResult object
-- Client-side validation (_TextAnalyzerConfigCreate, stopwords/stopword_presets mutex)
+- Client-side validation (TextAnalyzerConfigCreate, stopwords/stopword_presets mutex)
 - Version gate (>= 1.37.0)
 - Both sync and async client paths
 
@@ -20,15 +20,15 @@
 import pytest_asyncio
 
 import weaviate
-from weaviate.collections.classes.config import (
+from weaviate.classes.tokenization import (
+    StopwordsCreate,
     StopwordsPreset,
+    TextAnalyzerConfigCreate,
     Tokenization,
-    _StopwordsCreate,
-    _TextAnalyzerConfigCreate,
+    TokenizeResult,
 )
 from weaviate.config import AdditionalConfig
 from weaviate.exceptions import WeaviateUnsupportedFeatureError
-from weaviate.tokenization.models import TokenizeResult
 
 
 @pytest.fixture(scope="module")
@@ -140,7 +140,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "The quick brown fox",
-                    "analyzer_config": _TextAnalyzerConfigCreate(
+                    "analyzer_config": TextAnalyzerConfigCreate(
                         stopword_preset=StopwordsPreset.NONE
                     ),
                 },
@@ -150,7 +150,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "L'école est fermée",
-                    "analyzer_config": _TextAnalyzerConfigCreate(ascii_fold=True),
+                    "analyzer_config": TextAnalyzerConfigCreate(ascii_fold=True),
                 },
                 ["l", "ecole", "est", "fermee"],
                 ["l", "ecole", "est", "fermee"],
@@ -158,7 +158,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "L'école est fermée",
-                    "analyzer_config": _TextAnalyzerConfigCreate(
+                    "analyzer_config": TextAnalyzerConfigCreate(
                         ascii_fold=True, ascii_fold_ignore=["é"]
                     ),
                 },
@@ -168,9 +168,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "The quick brown fox",
-                    "analyzer_config": _TextAnalyzerConfigCreate(
-                        stopword_preset=StopwordsPreset.EN
-                    ),
+                    "analyzer_config": TextAnalyzerConfigCreate(stopword_preset=StopwordsPreset.EN),
                 },
                 ["the", "quick", "brown", "fox"],
                 ["quick", "brown", "fox"],
@@ -178,7 +176,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "The quick brown fox",
-                    "analyzer_config": _TextAnalyzerConfigCreate(stopword_preset="en"),
+                    "analyzer_config": TextAnalyzerConfigCreate(stopword_preset="en"),
                 },
                 ["the", "quick", "brown", "fox"],
                 ["quick", "brown", "fox"],
@@ -186,7 +184,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "The école est fermée",
-                    "analyzer_config": _TextAnalyzerConfigCreate(
+                    "analyzer_config": TextAnalyzerConfigCreate(
                         ascii_fold=True,
                         ascii_fold_ignore=["é"],
                         stopword_preset=StopwordsPreset.EN,
@@ -198,7 +196,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "the quick brown fox",
-                    "stopwords": _StopwordsCreate(
+                    "stopwords": StopwordsCreate(
                         preset=StopwordsPreset.EN, additions=["quick"], removals=None
                     ),
                 },
@@ -208,7 +206,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "the quick hello world",
-                    "stopwords": _StopwordsCreate(preset=None, additions=["hello"], removals=None),
+                    "stopwords": StopwordsCreate(preset=None, additions=["hello"], removals=None),
                 },
                 ["the", "quick", "hello", "world"],
                 ["quick", "world"],
@@ -216,7 +214,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "the quick is fast",
-                    "stopwords": _StopwordsCreate(preset=None, additions=None, removals=["the"]),
+                    "stopwords": StopwordsCreate(preset=None, additions=None, removals=["the"]),
                 },
                 ["the", "quick", "is", "fast"],
                 ["the", "quick", "fast"],
@@ -224,7 +222,7 @@ def test_tokenization_enum(
             (
                 {
                     "text": "hello world test",
-                    "analyzer_config": _TextAnalyzerConfigCreate(stopword_preset="custom"),
+                    "analyzer_config": TextAnalyzerConfigCreate(stopword_preset="custom"),
                     "stopword_presets": {"custom": ["test"]},
                 },
                 ["hello", "world", "test"],
@@ -276,7 +274,7 @@ def test_text_from_collection_config(
         result = client.tokenization.text(
             text="the quick brown fox",
             tokenization=recipe.tokenization,
-            stopwords=_StopwordsCreate(**stopwords.__dict__),
+            stopwords=StopwordsCreate(**stopwords.__dict__),
         )
         assert result.indexed == ["the", "quick", "brown", "fox"]
         assert result.query == ["brown", "fox"]
@@ -294,7 +292,7 @@ def test_property_and_generic_endpoints_agree(
         via_generic = client.tokenization.text(
             text=text,
             tokenization=recipe.tokenization,
-            stopwords=_StopwordsCreate(**stopwords.__dict__),
+            stopwords=StopwordsCreate(**stopwords.__dict__),
         )
 
         assert via_property.indexed == via_generic.indexed
@@ -360,7 +358,7 @@ class TestClientSideValidation:
     )
     def test_ascii_fold_ignore_without_fold_raises(self, kwargs: dict) -> None:
         with pytest.raises(ValueError, match="asciiFoldIgnore"):
-            _TextAnalyzerConfigCreate(**kwargs)
+            TextAnalyzerConfigCreate(**kwargs)
 
     @pytest.mark.parametrize(
         "kwargs,expected",
@@ -385,7 +383,7 @@ def test_ascii_fold_ignore_without_fold_raises(self, kwargs: dict) -> None:
         ids=["fold_with_ignore", "fold_without_ignore", "stopword_preset_only", "empty"],
     )
     def test_valid_config(self, kwargs: dict, expected: dict) -> None:
-        cfg = _TextAnalyzerConfigCreate(**kwargs)
+        cfg = TextAnalyzerConfigCreate(**kwargs)
         for attr, value in expected.items():
             assert getattr(cfg, attr) == value
 
@@ -397,9 +395,7 @@ def test_stopwords_and_stopword_presets_mutex(self, client: weaviate.WeaviateCli
             client.tokenization.text(
                 text="hello",
                 tokenization=Tokenization.WORD,
-                stopwords=_StopwordsCreate(
-                    preset=StopwordsPreset.EN, additions=None, removals=None
-                ),
+                stopwords=StopwordsCreate(preset=StopwordsPreset.EN, additions=None, removals=None),
                 stopword_presets={"custom": ["hello"]},
             )
 
@@ -482,7 +478,7 @@ async def test_text_tokenize(self, async_client: weaviate.WeaviateAsyncClient) -
     async def test_text_with_stopwords_fallback(
         self, async_client: weaviate.WeaviateAsyncClient
     ) -> None:
-        sw = _StopwordsCreate(preset=StopwordsPreset.EN, additions=["quick"], removals=None)
+        sw = StopwordsCreate(preset=StopwordsPreset.EN, additions=["quick"], removals=None)
         result = await async_client.tokenization.text(
             text="the quick brown fox",
             tokenization=Tokenization.WORD,
diff --git a/weaviate/classes/config.py b/weaviate/classes/config.py
index 868cd1c79..c154062d3 100644
--- a/weaviate/classes/config.py
+++ b/weaviate/classes/config.py
@@ -11,8 +11,10 @@
     ReferenceProperty,
     ReplicationDeletionStrategy,
     Rerankers,
+    StopwordsCreate,
     StopwordsPreset,
     TextAnalyzerConfig,
+    TextAnalyzerConfigCreate,
     Tokenization,
     VectorDistances,
 )
@@ -39,8 +41,10 @@
     "PQEncoderType",
     "ReferenceProperty",
     "Rerankers",
+    "StopwordsCreate",
     "StopwordsPreset",
     "TextAnalyzerConfig",
+    "TextAnalyzerConfigCreate",
     "Tokenization",
     "Vectorizers",
     "VectorDistances",
diff --git a/weaviate/classes/tokenization.py b/weaviate/classes/tokenization.py
new file mode 100644
index 000000000..ffb050614
--- /dev/null
+++ b/weaviate/classes/tokenization.py
@@ -0,0 +1,15 @@
+from weaviate.collections.classes.config import (
+    StopwordsCreate,
+    StopwordsPreset,
+    TextAnalyzerConfigCreate,
+    Tokenization,
+)
+from weaviate.tokenization.models import TokenizeResult
+
+__all__ = [
+    "StopwordsCreate",
+    "StopwordsPreset",
+    "TextAnalyzerConfigCreate",
+    "Tokenization",
+    "TokenizeResult",
+]
diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 6d60482a3..068399d70 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -1647,6 +1647,7 @@ class _StopwordsConfig(_ConfigBase):
 
 
 StopwordsConfig = _StopwordsConfig
+StopwordsCreate = _StopwordsCreate
 
 
 @dataclass
@@ -2224,6 +2225,9 @@ def _validate_ascii_fold_ignore(self) -> "_TextAnalyzerConfigCreate":
         return self
 
 
+TextAnalyzerConfigCreate = _TextAnalyzerConfigCreate
+
+
 class Property(_ConfigCreateModel):
     """This class defines the structure of a data property that a collection can have within Weaviate.
 
diff --git a/weaviate/tokenization/async_.pyi b/weaviate/tokenization/async_.pyi
index 11f4a13fc..59e815d87 100644
--- a/weaviate/tokenization/async_.pyi
+++ b/weaviate/tokenization/async_.pyi
@@ -1,9 +1,9 @@
 from typing import Dict, List, Optional
 
 from weaviate.collections.classes.config import (
+    StopwordsCreate,
+    TextAnalyzerConfigCreate,
     Tokenization,
-    _StopwordsCreate,
-    _TextAnalyzerConfigCreate,
 )
 from weaviate.connect.v4 import ConnectionAsync
 from weaviate.tokenization.models import TokenizeResult
@@ -16,7 +16,7 @@ class _TokenizationAsync(_TokenizationExecutor[ConnectionAsync]):
         text: str,
         tokenization: Tokenization,
         *,
-        analyzer_config: Optional[_TextAnalyzerConfigCreate] = None,
-        stopwords: Optional[_StopwordsCreate] = None,
+        analyzer_config: Optional[TextAnalyzerConfigCreate] = None,
+        stopwords: Optional[StopwordsCreate] = None,
         stopword_presets: Optional[Dict[str, List[str]]] = None,
     ) -> TokenizeResult: ...
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index 25b36e1d3..a3beffd44 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -5,9 +5,9 @@
 from httpx import Response
 
 from weaviate.collections.classes.config import (
+    StopwordsCreate,
+    TextAnalyzerConfigCreate,
     Tokenization,
-    _StopwordsCreate,
-    _TextAnalyzerConfigCreate,
 )
 from weaviate.connect import executor
 from weaviate.connect.v4 import ConnectionType, _ExpectedStatusCodes
@@ -32,8 +32,8 @@ def text(
         text: str,
         tokenization: Tokenization,
         *,
-        analyzer_config: Optional[_TextAnalyzerConfigCreate] = None,
-        stopwords: Optional[_StopwordsCreate] = None,
+        analyzer_config: Optional[TextAnalyzerConfigCreate] = None,
+        stopwords: Optional[StopwordsCreate] = None,
         stopword_presets: Optional[Dict[str, List[str]]] = None,
     ) -> executor.Result[TokenizeResult]:
         """Tokenize text using the generic /v1/tokenize endpoint.
diff --git a/weaviate/tokenization/sync.pyi b/weaviate/tokenization/sync.pyi
index d931aae51..2c2470f85 100644
--- a/weaviate/tokenization/sync.pyi
+++ b/weaviate/tokenization/sync.pyi
@@ -1,9 +1,9 @@
 from typing import Dict, List, Optional
 
 from weaviate.collections.classes.config import (
+    StopwordsCreate,
+    TextAnalyzerConfigCreate,
     Tokenization,
-    _StopwordsCreate,
-    _TextAnalyzerConfigCreate,
 )
 from weaviate.connect.v4 import ConnectionSync
 from weaviate.tokenization.models import TokenizeResult
@@ -16,7 +16,7 @@ class _Tokenization(_TokenizationExecutor[ConnectionSync]):
         text: str,
         tokenization: Tokenization,
         *,
-        analyzer_config: Optional[_TextAnalyzerConfigCreate] = None,
-        stopwords: Optional[_StopwordsCreate] = None,
+        analyzer_config: Optional[TextAnalyzerConfigCreate] = None,
+        stopwords: Optional[StopwordsCreate] = None,
         stopword_presets: Optional[Dict[str, List[str]]] = None,
     ) -> TokenizeResult: ...

From 55b136adfd37f8289b1aa9ffd3816335b25fd599 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 21 Apr 2026 15:40:47 +0200
Subject: [PATCH 94/99] Add overloads for exclusivity of stopwrods

---
 integration/test_tokenize.py      | 13 ++---
 weaviate/tokenization/async_.pyi  | 17 +++++--
 weaviate/tokenization/executor.py | 83 ++++++++++++++++++++++++-------
 weaviate/tokenization/sync.pyi    | 17 +++++--
 4 files changed, 93 insertions(+), 37 deletions(-)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index a5b16da32..dc244d2c3 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -274,7 +274,7 @@ def test_text_from_collection_config(
         result = client.tokenization.text(
             text="the quick brown fox",
             tokenization=recipe.tokenization,
-            stopwords=StopwordsCreate(**stopwords.__dict__),
+            stopwords=stopwords,
         )
         assert result.indexed == ["the", "quick", "brown", "fox"]
         assert result.query == ["brown", "fox"]
@@ -292,7 +292,7 @@ def test_property_and_generic_endpoints_agree(
         via_generic = client.tokenization.text(
             text=text,
             tokenization=recipe.tokenization,
-            stopwords=StopwordsCreate(**stopwords.__dict__),
+            stopwords=stopwords,
         )
 
         assert via_property.indexed == via_generic.indexed
@@ -308,13 +308,6 @@ def test_property_and_generic_endpoints_agree(
 class TestDeserialization:
     """Verify the client correctly deserializes response fields into TokenizeResult."""
 
-    def test_generic_result_shape(self, client: weaviate.WeaviateClient) -> None:
-        """Generic endpoint response deserializes into TokenizeResult with indexed and query lists."""
-        result = client.tokenization.text(text="hello", tokenization=Tokenization.WORD)
-        assert isinstance(result, TokenizeResult)
-        assert isinstance(result.indexed, list)
-        assert isinstance(result.query, list)
-
     def test_property_result_shape(self, client: weaviate.WeaviateClient) -> None:
         """Property endpoint response deserializes into TokenizeResult — server resolves tokenization from the property's schema."""
         client.collections.delete("TestDeserPropTypes")
@@ -405,7 +398,7 @@ def test_stopwords_and_stopword_presets_mutex(self, client: weaviate.WeaviateCli
             ({"custom": "hello"}, "must be a list of strings"),
             (
                 {
-                    "custom": _StopwordsCreate(
+                    "custom": StopwordsCreate(
                         preset=StopwordsPreset.EN, additions=None, removals=None
                     ),
                 },
diff --git a/weaviate/tokenization/async_.pyi b/weaviate/tokenization/async_.pyi
index 59e815d87..6bd2d9e8a 100644
--- a/weaviate/tokenization/async_.pyi
+++ b/weaviate/tokenization/async_.pyi
@@ -1,4 +1,4 @@
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, overload
 
 from weaviate.collections.classes.config import (
     StopwordsCreate,
@@ -11,12 +11,21 @@ from weaviate.tokenization.models import TokenizeResult
 from .executor import _TokenizationExecutor
 
 class _TokenizationAsync(_TokenizationExecutor[ConnectionAsync]):
+    @overload
     async def text(
         self,
         text: str,
         tokenization: Tokenization,
         *,
-        analyzer_config: Optional[TextAnalyzerConfigCreate] = None,
-        stopwords: Optional[StopwordsCreate] = None,
-        stopword_presets: Optional[Dict[str, List[str]]] = None,
+        analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
+        stopwords: Optional[StopwordsCreate] = ...,
+    ) -> TokenizeResult: ...
+    @overload
+    async def text(
+        self,
+        text: str,
+        tokenization: Tokenization,
+        *,
+        analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
+        stopword_presets: Optional[Dict[str, List[str]]] = ...,
     ) -> TokenizeResult: ...
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index a3beffd44..150cc6dd9 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -1,6 +1,6 @@
 """Tokenize executor."""
 
-from typing import Any, Dict, Generic, List, Optional
+from typing import Any, Dict, Generic, List, Optional, overload
 
 from httpx import Response
 
@@ -27,6 +27,29 @@ def __check_version(self) -> None:
                 "1.37.0",
             )
 
+    # Overloads make ``stopwords`` and ``stopword_presets`` mutually exclusive
+    # at type-check time. Passing both is additionally rejected at runtime with
+    # ``ValueError`` in the implementation below.
+    @overload
+    def text(
+        self,
+        text: str,
+        tokenization: Tokenization,
+        *,
+        analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
+        stopwords: Optional[StopwordsCreate] = ...,
+    ) -> executor.Result[TokenizeResult]: ...
+
+    @overload
+    def text(
+        self,
+        text: str,
+        tokenization: Tokenization,
+        *,
+        analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
+        stopword_presets: Optional[Dict[str, List[str]]] = ...,
+    ) -> executor.Result[TokenizeResult]: ...
+
     def text(
         self,
         text: str,
@@ -40,33 +63,55 @@ def text(
 
         For ``word`` tokenization the server defaults to the built-in ``en``
         stopword preset when no stopword configuration is supplied. Pass
-        ``analyzer_config=Configure.text_analyzer(stopword_preset=StopwordsPreset.NONE)``
-        (or equivalent) to opt out.
+        ``analyzer_config=TextAnalyzerConfigCreate(stopword_preset="none")``
+        or equivalent to opt out.
+
+        Call patterns for stopword handling (``stopwords`` and
+        ``stopword_presets`` are mutually exclusive — pass at most one):
+
+        1. **No stopword config** — rely on the server default (``en`` for
+           word tokenization, none otherwise)::
+
+               client.tokenization.text(text=..., tokenization=Tokenization.WORD)
+
+        2. **Apply a one-off stopwords block** via ``stopwords`` — the block
+           filters the query tokens directly, same shape as a collection's
+           ``invertedIndexConfig.stopwords``::
+
+               client.tokenization.text(
+                   text=...,
+                   tokenization=Tokenization.WORD,
+                   stopwords=StopwordsCreate(preset=StopwordsPreset.EN, additions=["foo"]),
+               )
+
+        3. **Register a named-preset catalog** via ``stopword_presets`` and
+           reference one by name from ``analyzer_config.stopword_preset``.
+           The catalog can also override built-in presets such as ``en``::
+
+               client.tokenization.text(
+                   text=...,
+                   tokenization=Tokenization.WORD,
+                   analyzer_config=TextAnalyzerConfigCreate(stopword_preset="custom"),
+                   stopword_presets={"custom": ["foo", "bar"]},
+               )
 
         Args:
             text: The text to tokenize.
-            tokenization: The tokenization method to use (e.g. Tokenization.WORD).
+            tokenization: The tokenization method to use (e.g. ``Tokenization.WORD``).
             analyzer_config: Text analyzer settings (ASCII folding, stopword
                 preset name), built via ``Configure.text_analyzer(...)``.
                 ``stopword_preset`` may reference a built-in preset
                 (``en`` / ``none``) or a name defined in ``stopword_presets``.
-            stopwords: Fallback stopword config applied when
-                ``analyzer_config.stopword_preset`` is not set. Same shape as a
-                collection's ``invertedIndexConfig.stopwords`` — a base preset
-                optionally tweaked with ``additions`` / ``removals``. An empty
-                ``preset`` defaults to ``en``.
-            stopword_presets: User-defined named stopword presets, each a plain
-                list of words. A name matching a built-in (``en`` / ``none``)
-                replaces the built-in entirely.
-
-        Note:
-            ``stopwords`` and ``stopword_presets`` are mutually exclusive on the
-            server — pass one or the other, not both. The server returns HTTP
-            422 if both are supplied.
+            stopwords: One-off stopwords block applied directly to this request.
+                Mutually exclusive with ``stopword_presets``.
+            stopword_presets: Named-preset catalog (name → word list). Entries
+                can be referenced from ``analyzer_config.stopword_preset`` or
+                override built-ins like ``en``. Mutually exclusive with
+                ``stopwords``.
 
         Returns:
-            A TokenizeResult with indexed and query token lists. The response
-            does not echo request fields back.
+            A ``TokenizeResult`` with indexed and query token lists. The generic
+            endpoint does not echo request fields back in the response.
 
         Raises:
             WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
diff --git a/weaviate/tokenization/sync.pyi b/weaviate/tokenization/sync.pyi
index 2c2470f85..7edf8994a 100644
--- a/weaviate/tokenization/sync.pyi
+++ b/weaviate/tokenization/sync.pyi
@@ -1,4 +1,4 @@
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, overload
 
 from weaviate.collections.classes.config import (
     StopwordsCreate,
@@ -11,12 +11,21 @@ from weaviate.tokenization.models import TokenizeResult
 from .executor import _TokenizationExecutor
 
 class _Tokenization(_TokenizationExecutor[ConnectionSync]):
+    @overload
     def text(
         self,
         text: str,
         tokenization: Tokenization,
         *,
-        analyzer_config: Optional[TextAnalyzerConfigCreate] = None,
-        stopwords: Optional[StopwordsCreate] = None,
-        stopword_presets: Optional[Dict[str, List[str]]] = None,
+        analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
+        stopwords: Optional[StopwordsCreate] = ...,
+    ) -> TokenizeResult: ...
+    @overload
+    def text(
+        self,
+        text: str,
+        tokenization: Tokenization,
+        *,
+        analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
+        stopword_presets: Optional[Dict[str, List[str]]] = ...,
     ) -> TokenizeResult: ...

From 7924e457cdd3315325db9a9e7ce2cc51bf7d6d04 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 21 Apr 2026 15:58:08 +0200
Subject: [PATCH 95/99] Accept collection config classes as stopwords

---
 weaviate/classes/tokenization.py       |  2 ++
 weaviate/collections/classes/config.py | 20 ++++++++++++++++++++
 weaviate/tokenization/async_.pyi       |  5 +++--
 weaviate/tokenization/executor.py      | 21 +++++++++++++++++----
 weaviate/tokenization/sync.pyi         |  5 +++--
 5 files changed, 45 insertions(+), 8 deletions(-)

diff --git a/weaviate/classes/tokenization.py b/weaviate/classes/tokenization.py
index ffb050614..0e89fc64b 100644
--- a/weaviate/classes/tokenization.py
+++ b/weaviate/classes/tokenization.py
@@ -1,4 +1,5 @@
 from weaviate.collections.classes.config import (
+    StopwordsConfig,
     StopwordsCreate,
     StopwordsPreset,
     TextAnalyzerConfigCreate,
@@ -7,6 +8,7 @@
 from weaviate.tokenization.models import TokenizeResult
 
 __all__ = [
+    "StopwordsConfig",
     "StopwordsCreate",
     "StopwordsPreset",
     "TextAnalyzerConfigCreate",
diff --git a/weaviate/collections/classes/config.py b/weaviate/collections/classes/config.py
index 068399d70..43d86375d 100644
--- a/weaviate/collections/classes/config.py
+++ b/weaviate/collections/classes/config.py
@@ -1,5 +1,6 @@
 import datetime
 from dataclasses import dataclass
+from dataclasses import fields as _dataclass_fields
 from typing import (
     Any,
     ClassVar,
@@ -1649,6 +1650,25 @@ class _StopwordsConfig(_ConfigBase):
 StopwordsConfig = _StopwordsConfig
 StopwordsCreate = _StopwordsCreate
 
+# Invariant: the read-side dataclass (_StopwordsConfig) and the write-side
+# pydantic model (_StopwordsCreate) must carry the same set of field names so
+# that values round-tripped from ``collection.config.get()`` can flow back into
+# ``tokenization.text()`` without silent data loss. If a field is added to one
+# but not the other, importing this module fails loudly; the read→write
+# conversion in ``weaviate/tokenization/executor.py::_TokenizationExecutor.text``
+# depends on this parity.
+_read_fields = {f.name for f in _dataclass_fields(_StopwordsConfig)}
+_write_fields = set(_StopwordsCreate.model_fields.keys())
+if _read_fields != _write_fields:
+    raise RuntimeError(
+        "_StopwordsConfig / _StopwordsCreate field drift detected — "
+        f"read-only={_read_fields - _write_fields}, "
+        f"write-only={_write_fields - _read_fields}. "
+        "Update both classes together, or adapt the read→write conversion in "
+        "weaviate/tokenization/executor.py::_TokenizationExecutor.text."
+    )
+del _read_fields, _write_fields
+
 
 @dataclass
 class _InvertedIndexConfig(_ConfigBase):
diff --git a/weaviate/tokenization/async_.pyi b/weaviate/tokenization/async_.pyi
index 6bd2d9e8a..156e25c90 100644
--- a/weaviate/tokenization/async_.pyi
+++ b/weaviate/tokenization/async_.pyi
@@ -1,6 +1,7 @@
-from typing import Dict, List, Optional, overload
+from typing import Dict, List, Optional, Union, overload
 
 from weaviate.collections.classes.config import (
+    StopwordsConfig,
     StopwordsCreate,
     TextAnalyzerConfigCreate,
     Tokenization,
@@ -18,7 +19,7 @@ class _TokenizationAsync(_TokenizationExecutor[ConnectionAsync]):
         tokenization: Tokenization,
         *,
         analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
-        stopwords: Optional[StopwordsCreate] = ...,
+        stopwords: Optional[Union[StopwordsCreate, StopwordsConfig]] = ...,
     ) -> TokenizeResult: ...
     @overload
     async def text(
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index 150cc6dd9..ea36e1cda 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -1,10 +1,11 @@
 """Tokenize executor."""
 
-from typing import Any, Dict, Generic, List, Optional, overload
+from typing import Any, Dict, Generic, List, Optional, Union, overload
 
 from httpx import Response
 
 from weaviate.collections.classes.config import (
+    StopwordsConfig,
     StopwordsCreate,
     TextAnalyzerConfigCreate,
     Tokenization,
@@ -29,7 +30,10 @@ def __check_version(self) -> None:
 
     # Overloads make ``stopwords`` and ``stopword_presets`` mutually exclusive
     # at type-check time. Passing both is additionally rejected at runtime with
-    # ``ValueError`` in the implementation below.
+    # ``ValueError`` in the implementation below. ``stopwords`` accepts either a
+    # ``StopwordsCreate`` (the write-side shape) or a ``StopwordsConfig`` (the
+    # read-side shape returned by ``collection.config.get()``), so values round-
+    # tripped through config reads can be passed back in directly.
     @overload
     def text(
         self,
@@ -37,7 +41,7 @@ def text(
         tokenization: Tokenization,
         *,
         analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
-        stopwords: Optional[StopwordsCreate] = ...,
+        stopwords: Optional[Union[StopwordsCreate, StopwordsConfig]] = ...,
     ) -> executor.Result[TokenizeResult]: ...
 
     @overload
@@ -56,7 +60,7 @@ def text(
         tokenization: Tokenization,
         *,
         analyzer_config: Optional[TextAnalyzerConfigCreate] = None,
-        stopwords: Optional[StopwordsCreate] = None,
+        stopwords: Optional[Union[StopwordsCreate, StopwordsConfig]] = None,
         stopword_presets: Optional[Dict[str, List[str]]] = None,
     ) -> executor.Result[TokenizeResult]:
         """Tokenize text using the generic /v1/tokenize endpoint.
@@ -134,6 +138,15 @@ def text(
                 payload["analyzerConfig"] = ac_dict
 
         if stopwords is not None:
+            if isinstance(stopwords, StopwordsConfig):
+                # Widen from the read-side shape returned by config.get() to the
+                # write-side shape the server expects. Field parity between the
+                # two classes is enforced at import time in
+                # ``weaviate/collections/classes/config.py``, so iterating
+                # ``StopwordsCreate.model_fields`` copies every field.
+                stopwords = StopwordsCreate(
+                    **{name: getattr(stopwords, name) for name in StopwordsCreate.model_fields}
+                )
             sw_dict = stopwords._to_dict()
             if sw_dict:
                 payload["stopwords"] = sw_dict
diff --git a/weaviate/tokenization/sync.pyi b/weaviate/tokenization/sync.pyi
index 7edf8994a..389edd485 100644
--- a/weaviate/tokenization/sync.pyi
+++ b/weaviate/tokenization/sync.pyi
@@ -1,6 +1,7 @@
-from typing import Dict, List, Optional, overload
+from typing import Dict, List, Optional, Union, overload
 
 from weaviate.collections.classes.config import (
+    StopwordsConfig,
     StopwordsCreate,
     TextAnalyzerConfigCreate,
     Tokenization,
@@ -18,7 +19,7 @@ class _Tokenization(_TokenizationExecutor[ConnectionSync]):
         tokenization: Tokenization,
         *,
         analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
-        stopwords: Optional[StopwordsCreate] = ...,
+        stopwords: Optional[Union[StopwordsCreate, StopwordsConfig]] = ...,
     ) -> TokenizeResult: ...
     @overload
     def text(

From 64bed62ea2dc6f3a05984d1ae4ce0700600027c7 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 21 Apr 2026 16:24:06 +0200
Subject: [PATCH 96/99] Improve docstring

---
 weaviate/tokenization/executor.py | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index ea36e1cda..0d287ba0e 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -107,11 +107,15 @@ def text(
                 ``stopword_preset`` may reference a built-in preset
                 (``en`` / ``none``) or a name defined in ``stopword_presets``.
             stopwords: One-off stopwords block applied directly to this request.
+                Mirrors the collection-level ``invertedIndexConfig.stopwords``
+                shape — hence the rich model with preset / additions / removals.
                 Mutually exclusive with ``stopword_presets``.
-            stopword_presets: Named-preset catalog (name → word list). Entries
-                can be referenced from ``analyzer_config.stopword_preset`` or
-                override built-ins like ``en``. Mutually exclusive with
-                ``stopwords``.
+            stopword_presets: Named-preset catalog (name → word list). Mirrors
+                the property-level preset catalog — a plain mapping, since a
+                property only references a preset by name (via
+                ``analyzer_config.stopword_preset``) rather than carrying the
+                full stopwords block. Entries can override built-ins like
+                ``en``. Mutually exclusive with ``stopwords``.
 
         Returns:
             A ``TokenizeResult`` with indexed and query token lists. The generic

From 220e839360848a8c67b7eab322a97b232d12d5c9 Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Tue, 21 Apr 2026 16:26:17 +0200
Subject: [PATCH 97/99] Hook up tokenization and clean up model

---
 weaviate/classes/__init__.py    | 2 ++
 weaviate/tokenization/models.py | 4 +---
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/weaviate/classes/__init__.py b/weaviate/classes/__init__.py
index d495744ac..69af5d920 100644
--- a/weaviate/classes/__init__.py
+++ b/weaviate/classes/__init__.py
@@ -13,6 +13,7 @@
     rbac,
     replication,
     tenants,
+    tokenization,
 )  # noqa: F401
 from .config import ConsistencyLevel
 
@@ -29,6 +30,7 @@
     "init",
     "query",
     "tenants",
+    "tokenization",
     "rbac",
     "replication",
 ]
diff --git a/weaviate/tokenization/models.py b/weaviate/tokenization/models.py
index f8fe7cb67..baeac140c 100644
--- a/weaviate/tokenization/models.py
+++ b/weaviate/tokenization/models.py
@@ -2,7 +2,7 @@
 
 from typing import List
 
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel
 
 
 class TokenizeResult(BaseModel):
@@ -13,7 +13,5 @@ class TokenizeResult(BaseModel):
         query: Tokens as they would be used for querying (after stopword removal).
     """
 
-    model_config = ConfigDict(populate_by_name=True, arbitrary_types_allowed=True)
-
     indexed: List[str]
     query: List[str]

From 081aaef36f83890eeb839e44363851477c44d1ca Mon Sep 17 00:00:00 2001
From: Dirk Kulawiak <dirk@semi.technology>
Date: Wed, 22 Apr 2026 16:54:36 +0200
Subject: [PATCH 98/99] Move property back to tokenization

---
 integration/test_tokenize.py            | 18 ++++++-----
 weaviate/collections/config/async_.pyi  |  2 --
 weaviate/collections/config/executor.py | 40 -------------------------
 weaviate/collections/config/sync.pyi    |  2 --
 weaviate/tokenization/async_.pyi        |  3 ++
 weaviate/tokenization/executor.py       | 40 +++++++++++++++++++++++++
 weaviate/tokenization/sync.pyi          |  1 +
 7 files changed, 54 insertions(+), 52 deletions(-)

diff --git a/integration/test_tokenize.py b/integration/test_tokenize.py
index dc244d2c3..d2d46916d 100644
--- a/integration/test_tokenize.py
+++ b/integration/test_tokenize.py
@@ -288,7 +288,9 @@ def test_property_and_generic_endpoints_agree(
         stopwords = config.inverted_index_config.stopwords
 
         text = "the quick brown fox"
-        via_property = recipe_collection.config.tokenize_property(property_name="recipe", text=text)
+        via_property = client.tokenization.for_property(
+            collection=recipe_collection.name, property_name="recipe", text=text
+        )
         via_generic = client.tokenization.text(
             text=text,
             tokenization=recipe.tokenization,
@@ -325,8 +327,9 @@ def test_property_result_shape(self, client: weaviate.WeaviateClient) -> None:
                     ],
                 }
             )
-            col = client.collections.get("TestDeserPropTypes")
-            result = col.config.tokenize_property(property_name="tag", text="  Hello World  ")
+            result = client.tokenization.for_property(
+                collection="TestDeserPropTypes", property_name="tag", text="  Hello World  "
+            )
             assert isinstance(result, TokenizeResult)
             assert result.indexed == ["Hello World"]
         finally:
@@ -442,9 +445,8 @@ def test_text_raises_on_old_server(self, client: weaviate.WeaviateClient) -> Non
     def test_tokenize_property_raises_on_old_server(self, client: weaviate.WeaviateClient) -> None:
         if client._connection._weaviate_version.is_at_least(1, 37, 0):
             pytest.skip("Version gate only applies to Weaviate < 1.37.0")
-        col = client.collections.get("Any")
         with pytest.raises(WeaviateUnsupportedFeatureError):
-            col.config.tokenize_property(property_name="title", text="hello")
+            client.tokenization.for_property(collection="Any", property_name="title", text="hello")
 
 
 # ---------------------------------------------------------------------------
@@ -454,7 +456,7 @@ def test_tokenize_property_raises_on_old_server(self, client: weaviate.WeaviateC
 
 @pytest.mark.usefixtures("require_1_37")
 class TestAsyncClient:
-    """Verify text() and tokenize_property() work through the async client."""
+    """Verify tokenization.text() and tokenization.for_property() work through the async client."""
 
     @pytest.mark.asyncio
     async def test_text_tokenize(self, async_client: weaviate.WeaviateAsyncClient) -> None:
@@ -498,8 +500,8 @@ async def test_property_tokenize(self, async_client: weaviate.WeaviateAsyncClien
                     ],
                 }
             )
-            col = async_client.collections.get("TestAsyncPropTokenize")
-            result = await col.config.tokenize_property(
+            result = await async_client.tokenization.for_property(
+                collection="TestAsyncPropTokenize",
                 property_name="title",
                 text="The quick brown fox",
             )
diff --git a/weaviate/collections/config/async_.pyi b/weaviate/collections/config/async_.pyi
index a1f740ded..015b70dab 100644
--- a/weaviate/collections/config/async_.pyi
+++ b/weaviate/collections/config/async_.pyi
@@ -27,7 +27,6 @@ from weaviate.collections.classes.config import (
 from weaviate.collections.classes.config_object_ttl import _ObjectTTLConfigUpdate
 from weaviate.collections.classes.config_vector_index import _VectorIndexConfigDynamicUpdate
 from weaviate.connect.v4 import ConnectionAsync
-from weaviate.tokenization.models import TokenizeResult
 
 from .executor import _ConfigCollectionExecutor
 
@@ -91,4 +90,3 @@ class _ConfigCollectionAsync(_ConfigCollectionExecutor[ConnectionAsync]):
         self, *, vector_config: Union[_VectorConfigCreate, List[_VectorConfigCreate]]
     ) -> None: ...
     async def delete_property_index(self, property_name: str, index_name: IndexName) -> bool: ...
-    async def tokenize_property(self, property_name: str, text: str) -> TokenizeResult: ...
diff --git a/weaviate/collections/config/executor.py b/weaviate/collections/config/executor.py
index fe9f5ec0d..103ab70ac 100644
--- a/weaviate/collections/config/executor.py
+++ b/weaviate/collections/config/executor.py
@@ -56,7 +56,6 @@
     WeaviateInvalidInputError,
     WeaviateUnsupportedFeatureError,
 )
-from weaviate.tokenization.models import TokenizeResult
 from weaviate.util import (
     _capitalize_first_letter,
     _decode_json_response_dict,
@@ -667,42 +666,3 @@ def resp(res: Response) -> bool:
             error_msg="Property may not exist",
             status_codes=_ExpectedStatusCodes(ok_in=[200], error="property exists"),
         )
-
-    def tokenize_property(
-        self,
-        property_name: str,
-        text: str,
-    ) -> executor.Result[TokenizeResult]:
-        """Tokenize text using a property's configured tokenization settings.
-
-        Args:
-            property_name: The property name whose tokenization config to use.
-            text: The text to tokenize.
-
-        Returns:
-            A TokenizeResult with indexed and query token lists.
-
-        Raises:
-            WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
-        """
-        if self._connection._weaviate_version.is_lower_than(1, 37, 0):
-            raise WeaviateUnsupportedFeatureError(
-                "Tokenization",
-                str(self._connection._weaviate_version),
-                "1.37.0",
-            )
-
-        path = f"/schema/{self._name}/properties/{property_name}/tokenize"
-        payload: Dict[str, Any] = {"text": text}
-
-        def resp(response: Response) -> TokenizeResult:
-            return TokenizeResult.model_validate(response.json())
-
-        return executor.execute(
-            response_callback=resp,
-            method=self._connection.post,
-            path=path,
-            weaviate_object=payload,
-            error_msg="Property tokenization failed",
-            status_codes=_ExpectedStatusCodes(ok_in=[200], error="tokenize property text"),
-        )
diff --git a/weaviate/collections/config/sync.pyi b/weaviate/collections/config/sync.pyi
index 3664a0e1b..e54d8c8fc 100644
--- a/weaviate/collections/config/sync.pyi
+++ b/weaviate/collections/config/sync.pyi
@@ -27,7 +27,6 @@ from weaviate.collections.classes.config import (
 from weaviate.collections.classes.config_object_ttl import _ObjectTTLConfigUpdate
 from weaviate.collections.classes.config_vector_index import _VectorIndexConfigDynamicUpdate
 from weaviate.connect.v4 import ConnectionSync
-from weaviate.tokenization.models import TokenizeResult
 
 from .executor import _ConfigCollectionExecutor
 
@@ -89,4 +88,3 @@ class _ConfigCollection(_ConfigCollectionExecutor[ConnectionSync]):
         self, *, vector_config: Union[_VectorConfigCreate, List[_VectorConfigCreate]]
     ) -> None: ...
     def delete_property_index(self, property_name: str, index_name: IndexName) -> bool: ...
-    def tokenize_property(self, property_name: str, text: str) -> TokenizeResult: ...
diff --git a/weaviate/tokenization/async_.pyi b/weaviate/tokenization/async_.pyi
index 156e25c90..ba12abc2a 100644
--- a/weaviate/tokenization/async_.pyi
+++ b/weaviate/tokenization/async_.pyi
@@ -30,3 +30,6 @@ class _TokenizationAsync(_TokenizationExecutor[ConnectionAsync]):
         analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
         stopword_presets: Optional[Dict[str, List[str]]] = ...,
     ) -> TokenizeResult: ...
+    async def for_property(
+        self, collection: str, property_name: str, text: str
+    ) -> TokenizeResult: ...
diff --git a/weaviate/tokenization/executor.py b/weaviate/tokenization/executor.py
index 0d287ba0e..33f1c05f9 100644
--- a/weaviate/tokenization/executor.py
+++ b/weaviate/tokenization/executor.py
@@ -14,6 +14,7 @@
 from weaviate.connect.v4 import ConnectionType, _ExpectedStatusCodes
 from weaviate.exceptions import WeaviateUnsupportedFeatureError
 from weaviate.tokenization.models import TokenizeResult
+from weaviate.util import _capitalize_first_letter
 
 
 class _TokenizationExecutor(Generic[ConnectionType]):
@@ -189,3 +190,42 @@ def resp(response: Response) -> TokenizeResult:
             error_msg="Tokenization failed",
             status_codes=_ExpectedStatusCodes(ok_in=[200], error="tokenize text"),
         )
+
+    def for_property(
+        self,
+        collection: str,
+        property_name: str,
+        text: str,
+    ) -> executor.Result[TokenizeResult]:
+        """Tokenize text using a property's configured tokenization settings.
+
+        The server resolves the tokenization and analyzer configuration from
+        the property's schema, so callers only supply the text.
+
+        Args:
+            collection: The collection that owns the property.
+            property_name: The property name whose tokenization config to use.
+            text: The text to tokenize.
+
+        Returns:
+            A TokenizeResult with indexed and query token lists.
+
+        Raises:
+            WeaviateUnsupportedFeatureError: If the server version is below 1.37.0.
+        """
+        self.__check_version()
+
+        path = f"/schema/{_capitalize_first_letter(collection)}/properties/{property_name}/tokenize"
+        payload: Dict[str, Any] = {"text": text}
+
+        def resp(response: Response) -> TokenizeResult:
+            return TokenizeResult.model_validate(response.json())
+
+        return executor.execute(
+            response_callback=resp,
+            method=self._connection.post,
+            path=path,
+            weaviate_object=payload,
+            error_msg="Property tokenization failed",
+            status_codes=_ExpectedStatusCodes(ok_in=[200], error="tokenize property text"),
+        )
diff --git a/weaviate/tokenization/sync.pyi b/weaviate/tokenization/sync.pyi
index 389edd485..71aaaea5c 100644
--- a/weaviate/tokenization/sync.pyi
+++ b/weaviate/tokenization/sync.pyi
@@ -30,3 +30,4 @@ class _Tokenization(_TokenizationExecutor[ConnectionSync]):
         analyzer_config: Optional[TextAnalyzerConfigCreate] = ...,
         stopword_presets: Optional[Dict[str, List[str]]] = ...,
     ) -> TokenizeResult: ...
+    def for_property(self, collection: str, property_name: str, text: str) -> TokenizeResult: ...

From 846344f5f2f541fc18eb23f3d013a74c1cdc078d Mon Sep 17 00:00:00 2001
From: Tommy Smith <tommy@weaviate.io>
Date: Thu, 23 Apr 2026 10:10:50 +0100
Subject: [PATCH 99/99] Add changelog

---
 docs/changelog.rst | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/docs/changelog.rst b/docs/changelog.rst
index 9ea29ab29..d84a58c83 100644
--- a/docs/changelog.rst
+++ b/docs/changelog.rst
@@ -1,6 +1,21 @@
 Changelog
 =========
 
+Version 4.21.0
+--------------
+This minor version includes:
+    - Support for new 1.37 features:
+        - Add support for the new ``blobHash`` property data type
+        - Add support for returning profiling when making queries with the ``return_metadata=["query_profile"]`` parameter
+        - Add support for on-demaned tokenization through the ``client.tokenize`` namespace
+        - Add support for managing permissions for accessing the native MCP server
+        - Add support for collection export
+        - Add support for incremental backups
+    - Minor bug fixes and improvements:
+        - Change ``alpha`` queries and aggregations to use server-side default parameter
+        - Fixes rare flakey behaviour of ``client.batch.stream`` on server hangup
+
+
 Version 4.20.5
 --------------
 This patch version includes: