Add voice-command router (injectable speech-to-text)

README.md

@@ -13,6 +13,7 @@
 
 ## Table of Contents
 
+- [What's new (2026-06-20) — Voice-Command Router](#whats-new-2026-06-20--voice-command-router)
 - [What's new (2026-06-20) — Locale-Aware Number, Currency & Date Parsing](#whats-new-2026-06-20--locale-aware-number-currency--date-parsing)
 - [What's new (2026-06-20) — Perceptual-Hash Image Dedupe](#whats-new-2026-06-20--perceptual-hash-image-dedupe)
 - [What's new (2026-06-20) — S3-Compatible Artifact Store](#whats-new-2026-06-20--s3-compatible-artifact-store)
@@ -96,6 +97,12 @@
 
 ---
 
+## What's new (2026-06-20) — Voice-Command Router
+
+Trigger flows hands-free from recognized speech. Full reference: [`docs/source/Eng/doc/new_features/v44_features_doc.rst`](docs/source/Eng/doc/new_features/v44_features_doc.rst).
+
+- **`VoiceRouter`** (`AC_voice_register` / `AC_voice_dispatch` / `AC_voice_list` / `AC_voice_clear`, `ac_*`): map spoken trigger phrases to `AC_*` action lists; feed it recognized text and it runs the closest registered command (phrase matching reuses the fuzzy matcher, so "save the file" fires "save file"). **Speech-to-text is out of scope and injectable** — the router takes text and a `recognizer`/`runner` callable, so routing is fully unit-tested without audio or any speech dependency (a real Vosk/mic recogniser plugs into `listen_once`).
+
 ## What's new (2026-06-20) — Locale-Aware Number, Currency & Date Parsing
 
 Parse localized numbers/currency/dates. Full reference: [`docs/source/Eng/doc/new_features/v43_features_doc.rst`](docs/source/Eng/doc/new_features/v43_features_doc.rst).

README/README_zh-CN.md

@@ -12,6 +12,7 @@
 
 ## 目录
 
+- [本次更新 (2026-06-20) — 语音指令路由器](#本次更新-2026-06-20--语音指令路由器)
 - [本次更新 (2026-06-20) — 区域设置感知的数字、货币与日期解析](#本次更新-2026-06-20--区域设置感知的数字货币与日期解析)
 - [本次更新 (2026-06-20) — 感知哈希图像去重](#本次更新-2026-06-20--感知哈希图像去重)
 - [本次更新 (2026-06-20) — S3 兼容成品存储](#本次更新-2026-06-20--s3-兼容成品存储)
@@ -95,6 +96,12 @@
 
 ---
 
+## 本次更新 (2026-06-20) — 语音指令路由器
+
+以已识别语音免手动触发流程。完整参考:[`docs/source/Zh/doc/new_features/v44_features_doc.rst`](../docs/source/Zh/doc/new_features/v44_features_doc.rst)。
+
+- **`VoiceRouter`**(`AC_voice_register` / `AC_voice_dispatch` / `AC_voice_list` / `AC_voice_clear`、`ac_*`):将语音触发短语映射到 `AC_*` 动作列表;喂入已识别文本即执行最接近的已注册指令(短语匹配重用模糊匹配器,因此「save the file」会触发「save file」)。**语音转文本不在范围内且可注入** —— 路由器接受文本与 `recognizer`/`runner` 可调用对象,因此路由在无音频、无任何语音依赖下完整单元测试(真实 Vosk/麦克风识别器接入 `listen_once`)。
+
 ## 本次更新 (2026-06-20) — 区域设置感知的数字、货币与日期解析
 
 解析本地化的数字/货币/日期。完整参考:[`docs/source/Zh/doc/new_features/v43_features_doc.rst`](../docs/source/Zh/doc/new_features/v43_features_doc.rst)。

README/README_zh-TW.md

@@ -12,6 +12,7 @@
 
 ## 目錄
 
+- [本次更新 (2026-06-20) — 語音指令路由器](#本次更新-2026-06-20--語音指令路由器)
 - [本次更新 (2026-06-20) — 區域設定感知的數字、貨幣與日期解析](#本次更新-2026-06-20--區域設定感知的數字貨幣與日期解析)
 - [本次更新 (2026-06-20) — 感知雜湊影像去重](#本次更新-2026-06-20--感知雜湊影像去重)
 - [本次更新 (2026-06-20) — S3 相容成品儲存](#本次更新-2026-06-20--s3-相容成品儲存)
@@ -95,6 +96,12 @@
 
 ---
 
+## 本次更新 (2026-06-20) — 語音指令路由器
+
+以已辨識語音免手動觸發流程。完整參考:[`docs/source/Zh/doc/new_features/v44_features_doc.rst`](../docs/source/Zh/doc/new_features/v44_features_doc.rst)。
+
+- **`VoiceRouter`**(`AC_voice_register` / `AC_voice_dispatch` / `AC_voice_list` / `AC_voice_clear`、`ac_*`):將語音觸發片語對應到 `AC_*` 動作清單;餵入已辨識文字即執行最接近的已註冊指令(片語比對重用模糊比對器,因此「save the file」會觸發「save file」)。**語音轉文字不在範圍內且可注入** —— 路由器接受文字與 `recognizer`/`runner` 可呼叫物件,因此路由在無音訊、無任何語音相依下完整單元測試(真實 Vosk/麥克風辨識器接入 `listen_once`)。
+
 ## 本次更新 (2026-06-20) — 區域設定感知的數字、貨幣與日期解析
 
 解析在地化的數字/貨幣/日期。完整參考:[`docs/source/Zh/doc/new_features/v43_features_doc.rst`](../docs/source/Zh/doc/new_features/v43_features_doc.rst)。

docs/source/Eng/doc/new_features/v44_features_doc.rst

@@ -0,0 +1,57 @@
+Voice-Command Router
+====================
+
+``VoiceRouter`` maps spoken trigger *phrases* to ``AC_*`` action lists: feed it
+the text of a recognized utterance and it runs the closest registered command —
+hands-free triggering of automation flows. Phrase matching reuses the project's
+fuzzy matcher, so "save the file" still fires a ``"save file"`` command despite
+recogniser noise.
+
+Speech-to-text is intentionally **out of scope and injectable**: the router takes
+already-recognised *text*. A real microphone/Vosk recogniser is supplied as a
+``recognizer`` callable to :meth:`VoiceRouter.listen_once`, which keeps the
+routing logic fully unit-testable without audio or any speech dependency. Imports
+no ``PySide6``.
+
+Headless API
+------------
+
+.. code-block:: python
+
+    from je_auto_control import VoiceRouter
+
+    router = VoiceRouter(threshold=0.7)
+    router.register("save file", [["AC_hotkey", {"keys": ["ctrl", "s"]}]])
+    router.register("close window", [["AC_close_window", {}]])
+
+    router.dispatch("save the file")     # fuzzy-matches -> runs the save actions
+
+    # with a real recogniser (any callable returning text):
+    def vosk_listen() -> str:
+        ...                              # capture audio, return transcript
+    router.listen_once(vosk_listen)
+
+``dispatch`` (and ``listen_once``) accept a ``runner`` to execute the action list
+— it defaults to the executor; inject a fake to test routing without running real
+automation. ``match`` returns the best ``VoiceCommand`` at or above ``threshold``
+(or ``None``); ``register`` replaces an existing phrase; ``phrases`` / ``clear``
+inspect and reset.
+
+Executor commands
+-----------------
+
+A module-level default router backs the executor/MCP surfaces:
+
+================================ ===================================================
+Command                          Effect
+================================ ===================================================
+``AC_voice_register``            Map a ``phrase`` to an ``actions`` list.
+``AC_voice_dispatch``            Run the command best matching recognized ``text``.
+``AC_voice_list``                List registered phrases.
+``AC_voice_clear``               Remove all registered commands.
+================================ ===================================================
+
+``actions`` accepts a list or a JSON-string list (so the visual builder works).
+The same operations are exposed as MCP tools (``ac_voice_register`` /
+``ac_voice_dispatch`` / ``ac_voice_list`` / ``ac_voice_clear``) and as Script
+Builder commands under **Agent**.

docs/source/Eng/eng_index.rst

@@ -66,6 +66,7 @@ Comprehensive guides for all AutoControl features.
    doc/new_features/v41_features_doc
    doc/new_features/v42_features_doc
    doc/new_features/v43_features_doc
+   doc/new_features/v44_features_doc
    doc/ocr_backends/ocr_backends_doc
    doc/observability/observability_doc
    doc/operations_layer/operations_layer_doc

docs/source/Zh/doc/new_features/v44_features_doc.rst

@@ -0,0 +1,51 @@
+語音指令路由器
+==============
+
+``VoiceRouter`` 將語音觸發*片語*對應到 ``AC_*`` 動作清單:餵入一段已辨識語句的文字,它
+就會執行最接近的已註冊指令 —— 免手動觸發自動化流程。片語比對重用本專案的模糊比對器,因
+此即使辨識有雜訊,「save the file」仍會觸發 ``"save file"`` 指令。
+
+語音轉文字刻意**不在範圍內且可注入**:路由器接受的是已辨識的*文字*。真實的麥克風/Vosk
+辨識器以 ``recognizer`` 可呼叫物件傳入 :meth:`VoiceRouter.listen_once`,如此路由邏輯可
+在無音訊、無任何語音相依的情況下完整單元測試。不匯入 ``PySide6``。
+
+無頭 API
+--------
+
+.. code-block:: python
+
+    from je_auto_control import VoiceRouter
+
+    router = VoiceRouter(threshold=0.7)
+    router.register("save file", [["AC_hotkey", {"keys": ["ctrl", "s"]}]])
+    router.register("close window", [["AC_close_window", {}]])
+
+    router.dispatch("save the file")     # 模糊比對 -> 執行儲存動作
+
+    # 搭配真實辨識器(任何回傳文字的可呼叫物件):
+    def vosk_listen() -> str:
+        ...                              # 擷取音訊、回傳逐字稿
+    router.listen_once(vosk_listen)
+
+``dispatch``(與 ``listen_once``)接受 ``runner`` 來執行動作清單 —— 預設為執行器;注入假
+物件即可在不執行真實自動化下測試路由。``match`` 回傳達到或高於 ``threshold`` 的最佳
+``VoiceCommand``(否則 ``None``);``register`` 會取代既有片語;``phrases`` / ``clear``
+則用於檢視與重置。
+
+執行器指令
+----------
+
+模組層級的預設路由器支撐 executor/MCP 介面:
+
+================================ ===================================================
+指令                             效果
+================================ ===================================================
+``AC_voice_register``            將 ``phrase`` 對應到 ``actions`` 清單。
+``AC_voice_dispatch``            執行最符合已辨識 ``text`` 的指令。
+``AC_voice_list``                列出已註冊片語。
+``AC_voice_clear``               移除所有已註冊指令。
+================================ ===================================================
+
+``actions`` 接受清單或 JSON 字串清單(因此視覺化建構器可用)。相同操作亦提供為 MCP 工具
+(``ac_voice_register`` / ``ac_voice_dispatch`` / ``ac_voice_list`` /
+``ac_voice_clear``),以及 Script Builder 中 **Agent** 分類下的指令。

docs/source/Zh/zh_index.rst

@@ -66,6 +66,7 @@ AutoControl 所有功能的完整使用指南。
    doc/new_features/v41_features_doc
    doc/new_features/v42_features_doc
    doc/new_features/v43_features_doc
+   doc/new_features/v44_features_doc
    doc/ocr_backends/ocr_backends_doc
    doc/observability/observability_doc
    doc/operations_layer/operations_layer_doc

je_auto_control/__init__.py

@@ -247,6 +247,10 @@
 from je_auto_control.utils.locale_parse import (
     format_currency, format_date, format_decimal, parse_decimal, parse_number,
 )
+# Voice-command router (injectable speech-to-text)
+from je_auto_control.utils.voice import (
+    VoiceCommand, VoiceRouter, default_voice_router,
+)
 # Background popup/interrupt watchdog (unattended automation)
 from je_auto_control.utils.watchdog import (
     PopupWatchdog, WatchdogRule, default_popup_watchdog,
@@ -700,6 +704,7 @@ def start_autocontrol_gui(*args, **kwargs):
     "images_similar",
     "format_currency", "format_date", "format_decimal", "parse_decimal",
     "parse_number",
+    "VoiceCommand", "VoiceRouter", "default_voice_router",
     # MCP server
     "AuditLogger", "HttpMCPServer", "MCPContent", "MCPPrompt",
     "MCPPromptArgument", "MCPResource", "MCPServer", "MCPTool",

je_auto_control/gui/script_builder/command_schema.py

@@ -994,6 +994,31 @@ def _add_misc_specs(specs: List[CommandSpec]) -> None:
         ),
         description="Format an ISO date string for a locale.",
     ))
+    specs.append(CommandSpec(
+        "AC_voice_register", "Agent", "Voice: Register Command",
+        fields=(
+            FieldSpec("phrase", FieldType.STRING, placeholder="save file"),
+            FieldSpec("actions", FieldType.STRING,
+                      placeholder='[["AC_hotkey", {"keys": ["ctrl", "s"]}]]'),
+        ),
+        description="Map a spoken phrase to an action list (JSON).",
+    ))
+    specs.append(CommandSpec(
+        "AC_voice_dispatch", "Agent", "Voice: Dispatch Text",
+        fields=(FieldSpec("text", FieldType.STRING,
+                          placeholder="save the file"),),
+        description="Run the command best matching recognized text.",
+    ))
+    specs.append(CommandSpec(
+        "AC_voice_list", "Agent", "Voice: List Commands",
+        fields=(),
+        description="List registered voice-command phrases.",
+    ))
+    specs.append(CommandSpec(
+        "AC_voice_clear", "Agent", "Voice: Clear Commands",
+        fields=(),
+        description="Remove all registered voice commands.",
+    ))
     specs.append(CommandSpec(
         "AC_generate_sop", "Report", "Generate SOP Document",
         fields=(

je_auto_control/utils/executor/action_executor.py

@@ -3168,6 +3168,33 @@ def _format_date(value: str, locale: str = "en_US",
     return {"text": format_date(value, locale, fmt)}
 
 
+def _voice_register(phrase: str, actions: Any) -> Dict[str, Any]:
+    """Adapter: register a voice command on the default router."""
+    from je_auto_control.utils.voice import default_voice_router
+    default_voice_router.register(phrase, _coerce_list(actions))
+    return {"phrases": default_voice_router.phrases()}
+
+
+def _voice_dispatch(text: str) -> Dict[str, Any]:
+    """Adapter: run the command best matching recognized ``text``."""
+    from je_auto_control.utils.voice import default_voice_router
+    outcome = default_voice_router.dispatch(text)
+    return {"matched": outcome["matched"], "phrase": outcome["phrase"]}
+
+
+def _voice_list() -> Dict[str, Any]:
+    """Adapter: list registered voice-command phrases."""
+    from je_auto_control.utils.voice import default_voice_router
+    return {"phrases": default_voice_router.phrases()}
+
+
+def _voice_clear() -> Dict[str, Any]:
+    """Adapter: clear all registered voice commands."""
+    from je_auto_control.utils.voice import default_voice_router
+    default_voice_router.clear()
+    return {"cleared": True}
+
+
 class Executor:
     """
     Executor
@@ -3434,6 +3461,10 @@ def __init__(self):
             "AC_format_decimal": _format_decimal,
             "AC_format_currency": _format_currency,
             "AC_format_date": _format_date,
+            "AC_voice_register": _voice_register,
+            "AC_voice_dispatch": _voice_dispatch,
+            "AC_voice_list": _voice_list,
+            "AC_voice_clear": _voice_clear,
             "AC_a11y_record_start": _a11y_record_start,
             "AC_a11y_record_stop": _a11y_record_stop,
             "AC_a11y_record_events": _a11y_record_events,

je_auto_control/utils/mcp_server/tools/_factories.py

@@ -3008,6 +3008,47 @@ def locale_tools() -> List[MCPTool]:
     ]
 
 
+def voice_tools() -> List[MCPTool]:
+    return [
+        MCPTool(
+            name="ac_voice_register",
+            description=("Register a voice command: a trigger 'phrase' and an "
+                         "'actions' list (AC_* steps) to run when recognized "
+                         "speech best-matches it. Returns {phrases}."),
+            input_schema=schema(
+                {"phrase": {"type": "string"},
+                 "actions": {"type": "array", "items": {"type": "object"}}},
+                ["phrase", "actions"]),
+            handler=h.voice_register,
+            annotations=SIDE_EFFECT_ONLY,
+        ),
+        MCPTool(
+            name="ac_voice_dispatch",
+            description=("Run the command whose phrase best matches recognized "
+                         "'text' (fuzzy). Returns {matched, phrase}."),
+            input_schema=schema({"text": {"type": "string"}}, ["text"]),
+            handler=h.voice_dispatch,
+            annotations=SIDE_EFFECT_ONLY,
+        ),
+        MCPTool(
+            name="ac_voice_list",
+            description="List registered voice-command phrases. Returns "
+                        "{phrases}.",
+            input_schema=schema({}),
+            handler=h.voice_list,
+            annotations=READ_ONLY,
+        ),
+        MCPTool(
+            name="ac_voice_clear",
+            description="Remove all registered voice commands. Returns "
+                        "{cleared}.",
+            input_schema=schema({}),
+            handler=h.voice_clear,
+            annotations=SIDE_EFFECT_ONLY,
+        ),
+    ]
+
+
 def unattended_tools() -> List[MCPTool]:
     return [
         MCPTool(
@@ -4069,7 +4110,7 @@ def media_assert_tools() -> List[MCPTool]:
     credential_lease_tools, egress_tools, approval_testing_tools,
     trajectory_eval_tools, compliance_tools, agent_trace_tools,
     video_report_tools, fuzzy_tools, artifact_store_tools, image_dedup_tools,
-    locale_tools,
+    locale_tools, voice_tools,
     screen_record_tools,
     process_and_shell_tools, remote_desktop_tools, gamepad_tools,
     usb_passthrough_tools, assertion_tools, data_source_tools,

je_auto_control/utils/mcp_server/tools/_handlers.py

@@ -1448,6 +1448,29 @@ def format_date(value, locale="en_US", fmt="medium"):
     return {"text": _fmt(value, locale, fmt)}
 
 
+def voice_register(phrase, actions):
+    from je_auto_control.utils.voice import default_voice_router
+    default_voice_router.register(phrase, actions)
+    return {"phrases": default_voice_router.phrases()}
+
+
+def voice_dispatch(text):
+    from je_auto_control.utils.voice import default_voice_router
+    outcome = default_voice_router.dispatch(text)
+    return {"matched": outcome["matched"], "phrase": outcome["phrase"]}
+
+
+def voice_list():
+    from je_auto_control.utils.voice import default_voice_router
+    return {"phrases": default_voice_router.phrases()}
+
+
+def voice_clear():
+    from je_auto_control.utils.voice import default_voice_router
+    default_voice_router.clear()
+    return {"cleared": True}
+
+
 def vlm_locate(description: str,
                screen_region: Optional[List[int]] = None,
                model: Optional[str] = None) -> Optional[List[int]]:

je_auto_control/utils/voice/__init__.py

@@ -0,0 +1,6 @@
+"""Voice-command router: map recognized phrases to AC_* action lists."""
+from je_auto_control.utils.voice.voice_router import (
+    VoiceCommand, VoiceRouter, default_voice_router,
+)
+
+__all__ = ["VoiceCommand", "VoiceRouter", "default_voice_router"]