diff --git a/README.md b/README.md
index 039a1e42..7b2a179c 100644
--- a/README.md
+++ b/README.md
@@ -101,24 +101,24 @@ import { Factuality } from "autoevals";
 
 ## Using other AI providers
 
-When you use Autoevals, it will look for an `OPENAI_BASE_URL` environment variable to use as the base for requests to an OpenAI compatible API. If `OPENAI_BASE_URL` is not set, it will default to the [AI proxy](https://www.braintrust.dev/docs/guides/proxy).
+When you use Autoevals, it will look for an `OPENAI_BASE_URL` environment variable to use as the base for requests to an OpenAI-compatible API. If `OPENAI_BASE_URL` is not set, it will look for a `BRAINTRUST_AI_GATEWAY_URL` environment variable and then default to the [Braintrust Gateway](https://www.braintrust.dev/docs/deploy/gateway).
 
-If you choose to use the proxy, you'll also get:
+When you use the Braintrust Gateway, you'll also get:
 
 - Simplified access to many AI providers
 - Reduced costs with automatic request caching
 - Increased observability when you enable logging to Braintrust
 
-The proxy is free to use, even if you don't have a Braintrust account.
+The Braintrust-hosted Gateway is free to use while it is in beta.
 
-If you have a Braintrust account, you can optionally set the `BRAINTRUST_API_KEY` environment variable instead of `OPENAI_API_KEY` to unlock additional features like logging and monitoring. You can also route requests to [supported AI providers and models](https://www.braintrust.dev/docs/guides/proxy#supported-models) or custom models you have configured in Braintrust.
+Set the `BRAINTRUST_API_KEY` environment variable to authenticate Gateway requests. You can also route requests to supported AI providers and models or custom models you have configured in Braintrust.
 
 <div className="tabs">
 
 ### Python
 
 ```python
-# NOTE: ensure BRAINTRUST_API_KEY is set in your environment and OPENAI_API_KEY is not set
+# NOTE: ensure BRAINTRUST_API_KEY is set in your environment
 from autoevals.llm import *
 
 # Create an LLM-based evaluator using the Claude 3.5 Sonnet model from Anthropic
@@ -139,7 +139,7 @@ print(f"Factuality metadata: {result.metadata['rationale']}")
 ### TypeScript
 
 ```typescript
-// NOTE: ensure BRAINTRUST_API_KEY is set in your environment and OPENAI_API_KEY is not set
+// NOTE: ensure BRAINTRUST_API_KEY is set in your environment
 import { Factuality } from "autoevals";
 
 (async () => {
diff --git a/js/oai.test.ts b/js/oai.test.ts
index 20979a54..5adf9ba7 100644
--- a/js/oai.test.ts
+++ b/js/oai.test.ts
@@ -26,10 +26,14 @@ beforeAll(() => {
 
 let OPENAI_API_KEY: string | undefined;
 let OPENAI_BASE_URL: string | undefined;
+let BRAINTRUST_API_KEY: string | undefined;
+let BRAINTRUST_AI_GATEWAY_URL: string | undefined;
 
 beforeEach(() => {
   OPENAI_API_KEY = process.env.OPENAI_API_KEY;
   OPENAI_BASE_URL = process.env.OPENAI_BASE_URL;
+  BRAINTRUST_API_KEY = process.env.BRAINTRUST_API_KEY;
+  BRAINTRUST_AI_GATEWAY_URL = process.env.BRAINTRUST_AI_GATEWAY_URL;
 });
 
 afterEach(() => {
@@ -37,6 +41,8 @@ afterEach(() => {
 
   process.env.OPENAI_API_KEY = OPENAI_API_KEY;
   process.env.OPENAI_BASE_URL = OPENAI_BASE_URL;
+  process.env.BRAINTRUST_API_KEY = BRAINTRUST_API_KEY;
+  process.env.BRAINTRUST_AI_GATEWAY_URL = BRAINTRUST_AI_GATEWAY_URL;
 
   // Reset init state
   init({ client: undefined, defaultModel: undefined });
@@ -120,13 +126,14 @@ describe("OAI", () => {
     );
   });
 
-  test("calls proxy if everything unset", async () => {
+  test("calls gateway if everything unset", async () => {
     delete process.env.OPENAI_API_KEY;
     delete process.env.OPENAI_BASE_URL;
+    delete process.env.BRAINTRUST_AI_GATEWAY_URL;
+    process.env.BRAINTRUST_API_KEY = "braintrust-test-key";
 
     server.use(
-      http.post("https://api.braintrust.dev/v1/proxy/chat/completions", () => {
-        debugger;
+      http.post("https://gateway.braintrust.dev/chat/completions", () => {
         return HttpResponse.json(MOCK_OPENAI_COMPLETION_RESPONSE);
       }),
     );
@@ -137,7 +144,28 @@ describe("OAI", () => {
       messages: [{ role: "user", content: "Hello" }],
     });
 
-    debugger;
+    expect(response.choices[0].message.content).toBe(
+      "Hello, I am a mock response!",
+    );
+  });
+
+  test("uses configured Braintrust Gateway URL", async () => {
+    delete process.env.OPENAI_API_KEY;
+    delete process.env.OPENAI_BASE_URL;
+    process.env.BRAINTRUST_API_KEY = "braintrust-test-key";
+    process.env.BRAINTRUST_AI_GATEWAY_URL = "  https://gateway.example.com  ";
+
+    server.use(
+      http.post("https://gateway.example.com/chat/completions", () => {
+        return HttpResponse.json(MOCK_OPENAI_COMPLETION_RESPONSE);
+      }),
+    );
+
+    const client = buildOpenAIClient({});
+    const response = await client.chat.completions.create({
+      model: "gpt-4",
+      messages: [{ role: "user", content: "Hello" }],
+    });
 
     expect(response.choices[0].message.content).toBe(
       "Hello, I am a mock response!",
@@ -147,9 +175,11 @@ describe("OAI", () => {
   test("default wraps", async () => {
     delete process.env.OPENAI_API_KEY;
     delete process.env.OPENAI_BASE_URL;
+    delete process.env.BRAINTRUST_AI_GATEWAY_URL;
+    process.env.BRAINTRUST_API_KEY = "braintrust-test-key";
 
     server.use(
-      http.post("https://api.braintrust.dev/v1/proxy/chat/completions", () => {
+      http.post("https://gateway.braintrust.dev/chat/completions", () => {
         return HttpResponse.json(MOCK_OPENAI_COMPLETION_RESPONSE);
       }),
     );
@@ -173,9 +203,11 @@ describe("OAI", () => {
   test("wraps once", async () => {
     delete process.env.OPENAI_API_KEY;
     delete process.env.OPENAI_BASE_URL;
+    delete process.env.BRAINTRUST_AI_GATEWAY_URL;
+    process.env.BRAINTRUST_API_KEY = "braintrust-test-key";
 
     server.use(
-      http.post("https://api.braintrust.dev/v1/proxy/chat/completions", () => {
+      http.post("https://gateway.braintrust.dev/chat/completions", () => {
         return HttpResponse.json(MOCK_OPENAI_COMPLETION_RESPONSE);
       }),
     );
diff --git a/js/oai.ts b/js/oai.ts
index 15157180..6da1dee0 100644
--- a/js/oai.ts
+++ b/js/oai.ts
@@ -88,7 +88,18 @@ export function extractOpenAIArgs<T extends Record<string, unknown>>(
       };
 }
 
-const PROXY_URL = "https://api.braintrust.dev/v1/proxy";
+const DEFAULT_GATEWAY_URL = "https://gateway.braintrust.dev";
+
+const getGatewayURL = (): string =>
+  (process.env.BRAINTRUST_AI_GATEWAY_URL ?? "").trim() || DEFAULT_GATEWAY_URL;
+
+const isGatewayBaseURL = (baseURL: string): boolean => {
+  const normalizedBaseURL = baseURL.replace(/\/+$/, "");
+  return (
+    normalizedBaseURL === DEFAULT_GATEWAY_URL ||
+    normalizedBaseURL === getGatewayURL().replace(/\/+$/, "")
+  );
+};
 
 const resolveOpenAIClient = (options: OpenAIAuth): OpenAI => {
   const {
@@ -121,13 +132,18 @@ const resolveOpenAIClient = (options: OpenAIAuth): OpenAI => {
     });
   }
 
+  const baseURL =
+    openAiBaseUrl || process.env.OPENAI_BASE_URL || getGatewayURL();
+  const apiKey =
+    openAiApiKey ||
+    (isGatewayBaseURL(baseURL)
+      ? process.env.BRAINTRUST_API_KEY || process.env.OPENAI_API_KEY
+      : process.env.OPENAI_API_KEY || process.env.BRAINTRUST_API_KEY);
+
   return new OpenAI({
-    apiKey:
-      openAiApiKey ||
-      process.env.OPENAI_API_KEY ||
-      process.env.BRAINTRUST_API_KEY,
+    apiKey,
     organization: openAiOrganizationId,
-    baseURL: openAiBaseUrl || process.env.OPENAI_BASE_URL || PROXY_URL,
+    baseURL,
     defaultHeaders: openAiDefaultHeaders,
     dangerouslyAllowBrowser: openAiDangerouslyAllowBrowser,
   });
@@ -179,7 +195,7 @@ export interface InitOptions {
   /**
    * An OpenAI-compatible client to use for all evaluations.
    * This can be an OpenAI client, or any client that implements the OpenAI API
-   * (e.g., configured to use the Braintrust proxy with Anthropic, Gemini, etc.)
+   * (e.g., configured to use the Braintrust Gateway with Anthropic, Gemini, etc.)
    */
   client?: OpenAI;
   /**
@@ -192,7 +208,7 @@ export interface InitOptions {
    *   default models for different evaluation types. Only the specified models
    *   are updated; others remain unchanged.
    *
-   * When using non-OpenAI providers via the Braintrust proxy, set this to
+   * When using non-OpenAI providers via the Braintrust Gateway, set this to
    * the appropriate model string (e.g., "claude-3-5-sonnet-20241022").
    *
    * @example
@@ -243,14 +259,14 @@ export interface InitOptions {
  * init({ client: new OpenAI() });
  *
  * @example
- * // Using with Anthropic via Braintrust proxy
+ * // Using with Anthropic via Braintrust Gateway
  * import { init } from "autoevals";
  * import { OpenAI } from "openai";
  *
  * init({
  *   client: new OpenAI({
  *     apiKey: process.env.BRAINTRUST_API_KEY,
- *     baseURL: "https://api.braintrust.dev/v1/proxy",
+ *     baseURL: process.env.BRAINTRUST_AI_GATEWAY_URL || "https://gateway.braintrust.dev",
  *   }),
  *   defaultModel: {
  *     completion: "claude-3-5-sonnet-20241022",
diff --git a/py/autoevals/__init__.py b/py/autoevals/__init__.py
index 10bda72b..0a5841ad 100644
--- a/py/autoevals/__init__.py
+++ b/py/autoevals/__init__.py
@@ -15,7 +15,7 @@
 - Both sync and async evaluation support
 - Configurable scoring parameters
 - Detailed feedback through metadata
-- Integration with OpenAI and other LLM providers through Braintrust AI Proxy
+- Integration with OpenAI and other LLM providers through the Braintrust Gateway
 
 **Client setup**:
 
@@ -43,10 +43,10 @@
 evaluator = ClosedQA(client=client)
 ```
 
-**Multi-provider support via the Braintrust AI Proxy**:
+**Multi-provider support via the Braintrust Gateway**:
 
-Autoevals supports multiple LLM providers (Anthropic, Azure, etc.) through the Braintrust AI Proxy.
-Configure your client to use the proxy and set the default model:
+Autoevals supports multiple LLM providers (Anthropic, Azure, etc.) through the Braintrust Gateway.
+Configure your client to use the Gateway and set the default model:
 
 ```python
 import os
@@ -54,9 +54,9 @@
 from autoevals import init
 from autoevals.llm import Factuality
 
-# Configure client to use Braintrust AI Proxy with Claude
+# Configure client to use the Braintrust Gateway with Claude
 client = AsyncOpenAI(
-    base_url="https://api.braintrust.dev/v1/proxy",
+    base_url=os.getenv("BRAINTRUST_AI_GATEWAY_URL") or "https://gateway.braintrust.dev",
     api_key=os.getenv("BRAINTRUST_API_KEY"),
 )
 
diff --git a/py/autoevals/oai.py b/py/autoevals/oai.py
index 2ccb50f0..7a20e299 100644
--- a/py/autoevals/oai.py
+++ b/py/autoevals/oai.py
@@ -9,7 +9,16 @@
 from dataclasses import dataclass
 from typing import Any, Optional, Protocol, TypedDict, TypeVar, Union, cast, runtime_checkable
 
-PROXY_URL = "https://api.braintrust.dev/v1/proxy"
+GATEWAY_URL = "https://gateway.braintrust.dev"
+
+
+def _gateway_url() -> str:
+    return os.environ.get("BRAINTRUST_AI_GATEWAY_URL", "").strip() or GATEWAY_URL
+
+
+def _is_gateway_url(base_url: str) -> bool:
+    normalized_base_url = base_url.rstrip("/")
+    return normalized_base_url == GATEWAY_URL or normalized_base_url == _gateway_url().rstrip("/")
 
 
 class DefaultModelConfig(TypedDict, total=False):
@@ -195,7 +204,9 @@ def __post_init__(self):
         if not has_customization and not isinstance(self.openai, NamedWrapper):
             self.openai = wrap_openai(self.openai)
 
-        self._is_wrapped = isinstance(self.openai, NamedWrapper)
+        self._is_wrapped = isinstance(self.openai, NamedWrapper) or (
+            not has_customization and wrap_openai.__module__.startswith("braintrust.")
+        )
 
         openai_module = get_openai_module()
 
@@ -431,7 +442,7 @@ def init(
               models for different evaluation types. Only the specified models are updated;
               others remain unchanged.
 
-            When using non-OpenAI providers via the Braintrust proxy, set this to the
+            When using non-OpenAI providers via the Braintrust Gateway, set this to the
             appropriate model string (e.g., "claude-3-5-sonnet-20241022").
 
     Example:
@@ -448,7 +459,7 @@ def init(
             init(
                 client=OpenAI(
                     api_key=os.environ["BRAINTRUST_API_KEY"],
-                    base_url="https://api.braintrust.dev/v1/proxy",
+                    base_url=os.getenv("BRAINTRUST_AI_GATEWAY_URL") or "https://gateway.braintrust.dev",
                 ),
                 default_model={
                     "completion": "claude-3-5-sonnet-20241022",
@@ -520,7 +531,8 @@ def prepare_openai(
             Deprecated: Use the `client` argument and set the `openai`.
 
         base_url (str, optional): Base URL for API requests. If not provided, will
-            use OPENAI_BASE_URL from environment or fall back to PROXY_URL.
+            use OPENAI_BASE_URL from environment or fall back to BRAINTRUST_AI_GATEWAY_URL
+            or GATEWAY_URL.
             Deprecated: Use the `client` argument and set the `openai`.
 
     Returns:
@@ -559,11 +571,14 @@ def prepare_openai(
         )
         warned_deprecated_api_key_base_url = True
 
+    if base_url is None:
+        base_url = os.environ.get("OPENAI_BASE_URL") or _gateway_url()
     # prepare the default openai sdk, if not provided
     if api_key is None:
-        api_key = os.environ.get("OPENAI_API_KEY") or os.environ.get("BRAINTRUST_API_KEY")
-    if base_url is None:
-        base_url = os.environ.get("OPENAI_BASE_URL", PROXY_URL)
+        if _is_gateway_url(base_url):
+            api_key = os.environ.get("BRAINTRUST_API_KEY") or os.environ.get("OPENAI_API_KEY")
+        else:
+            api_key = os.environ.get("OPENAI_API_KEY") or os.environ.get("BRAINTRUST_API_KEY")
 
     if hasattr(openai_module, "OpenAI"):
         openai_module = cast(OpenAIV1Module, openai_module)
diff --git a/py/autoevals/test_oai.py b/py/autoevals/test_oai.py
index b3454dd7..1f3447d5 100644
--- a/py/autoevals/test_oai.py
+++ b/py/autoevals/test_oai.py
@@ -7,6 +7,7 @@
 
 from autoevals import init  # type: ignore[import]
 from autoevals.oai import (  # type: ignore[import]
+    GATEWAY_URL,
     LLMClient,
     OpenAIV0Module,
     OpenAIV1Module,
@@ -28,6 +29,7 @@ def reset_env_and_client(monkeypatch: pytest.MonkeyPatch):
     monkeypatch.delenv("OPENAI_API_KEY", raising=False)
     monkeypatch.setenv("OPENAI_API_KEY", "test-key")
     monkeypatch.setenv("OPENAI_BASE_URL", "http://test-url")
+    monkeypatch.delenv("BRAINTRUST_AI_GATEWAY_URL", raising=False)
     monkeypatch.setattr("autoevals.oai._named_wrapper", None)
     monkeypatch.setattr("autoevals.oai._wrap_openai", None)
     monkeypatch.setattr("autoevals.oai._openai_module", None)
@@ -89,6 +91,32 @@ def test_prepare_openai_defaults():
     assert openai_obj.base_url == "http://test-url"
 
 
+def test_prepare_openai_defaults_to_gateway(monkeypatch: pytest.MonkeyPatch):
+    monkeypatch.delenv("OPENAI_BASE_URL", raising=False)
+    monkeypatch.delenv("BRAINTRUST_AI_GATEWAY_URL", raising=False)
+    monkeypatch.setenv("OPENAI_API_KEY", "openai-key")
+    monkeypatch.setenv("BRAINTRUST_API_KEY", "braintrust-key")
+
+    prepared_client = prepare_openai()
+
+    openai_obj = unwrap_named_wrapper(prepared_client.openai)
+    assert openai_obj.api_key == "braintrust-key"
+    assert str(openai_obj.base_url).rstrip("/") == GATEWAY_URL
+
+
+def test_prepare_openai_uses_configured_gateway_url(monkeypatch: pytest.MonkeyPatch):
+    monkeypatch.delenv("OPENAI_BASE_URL", raising=False)
+    monkeypatch.setenv("BRAINTRUST_AI_GATEWAY_URL", "  https://gateway.example.com  ")
+    monkeypatch.setenv("OPENAI_API_KEY", "openai-key")
+    monkeypatch.setenv("BRAINTRUST_API_KEY", "braintrust-key")
+
+    prepared_client = prepare_openai()
+
+    openai_obj = unwrap_named_wrapper(prepared_client.openai)
+    assert openai_obj.api_key == "braintrust-key"
+    assert str(openai_obj.base_url).rstrip("/") == "https://gateway.example.com"
+
+
 def test_prepare_openai_with_plain_openai():
     client = openai.OpenAI(api_key="api-key", base_url="http://test")
     prepared_client = prepare_openai(client=client)