feat: TTS speed support (#321)

speed param (0.25-4.0). Edge->SSML, OpenAI->native, MiniMax->passthrough. 4 tests pass.

tests/tools/test_tts_speed.py

@@ -0,0 +1,20 @@
+"""Tests for TTS speed support (#321)."""
+import json
+from unittest.mock import patch
+class TestSchema:
+    def test_in(self):
+        from tools.tts_tool import TTS_SCHEMA
+        assert "speed" in TTS_SCHEMA["parameters"]["properties"]
+    def test_opt(self):
+        from tools.tts_tool import TTS_SCHEMA
+        assert "speed" not in TTS_SCHEMA["parameters"].get("required", [])
+class TestSig:
+    def test_has(self):
+        from tools.tts_tool import text_to_speech_tool
+        import inspect
+        assert "speed" in inspect.signature(text_to_speech_tool).parameters
+class TestRate:
+    def test_edge(self):
+        for s,e in [(1.0,"+0%"),(1.5,"+50%"),(0.5,"-50%")]:
+            p=int((s-1.0)*100)
+            assert (f"+{p}%" if p>=0 else f"{p}%")==e

tools/tts_tool.py

@@ -179,8 +179,10 @@ async def _generate_edge_tts(text: str, output_path: str, tts_config: Dict[str,
     _edge_tts = _import_edge_tts()
     edge_config = tts_config.get("edge", {})
     voice = edge_config.get("voice", DEFAULT_EDGE_VOICE)
-
-    communicate = _edge_tts.Communicate(text, voice)
+    speed = tts_config.get("_speed_override") or edge_config.get("speed", 1.0)
+    rate_pct = int((speed - 1.0) * 100)
+    rate_str = f"+{rate_pct}%" if rate_pct >= 0 else f"{rate_pct}%"
+    communicate = _edge_tts.Communicate(text, voice, rate=rate_str)
     await communicate.save(output_path)
     return output_path
 
@@ -262,11 +264,14 @@ def _generate_openai_tts(text: str, output_path: str, tts_config: Dict[str, Any]
     OpenAIClient = _import_openai_client()
     client = OpenAIClient(api_key=api_key, base_url=base_url)
     try:
+        speed = tts_config.get("_speed_override") or oai_config.get("speed", 1.0)
+        speed = max(0.25, min(4.0, speed))
         response = client.audio.speech.create(
             model=model,
             voice=voice,
             input=text,
             response_format=response_format,
+            speed=speed,
             extra_headers={"x-idempotency-key": str(uuid.uuid4())},
         )
 
@@ -305,7 +310,7 @@ def _generate_minimax_tts(text: str, output_path: str, tts_config: Dict[str, Any
     mm_config = tts_config.get("minimax", {})
     model = mm_config.get("model", DEFAULT_MINIMAX_MODEL)
     voice_id = mm_config.get("voice_id", DEFAULT_MINIMAX_VOICE_ID)
-    speed = mm_config.get("speed", 1)
+    speed = tts_config.get("_speed_override") or mm_config.get("speed", 1)
     vol = mm_config.get("vol", 1)
     pitch = mm_config.get("pitch", 0)
     base_url = mm_config.get("base_url", DEFAULT_MINIMAX_BASE_URL)
@@ -447,6 +452,7 @@ def _generate_neutts(text: str, output_path: str, tts_config: Dict[str, Any]) ->
 def text_to_speech_tool(
     text: str,
     output_path: Optional[str] = None,
+    speed: Optional[float] = None,
 ) -> str:
     """
     Convert text to speech audio.
@@ -474,6 +480,9 @@ def text_to_speech_tool(
         text = text[:MAX_TEXT_LENGTH]
 
     tts_config = _load_tts_config()
+    if speed is not None:
+        speed = max(0.25, min(4.0, speed))
+        tts_config["_speed_override"] = speed
     provider = _get_provider(tts_config)
 
     # Detect platform from gateway env var to choose the best output format.
@@ -966,6 +975,10 @@ TTS_SCHEMA = {
             "output_path": {
                 "type": "string",
                 "description": "Optional custom file path to save the audio. Defaults to ~/.hermes/audio_cache/<timestamp>.mp3"
+            },
+            "speed": {
+                "type": "number",
+                "description": "Speech speed multiplier. 1.0 = normal, 0.5 = half speed, 2.0 = double. Range: 0.25-4.0."
             }
         },
         "required": ["text"]
@@ -978,7 +991,8 @@ registry.register(
     schema=TTS_SCHEMA,
     handler=lambda args, **kw: text_to_speech_tool(
         text=args.get("text", ""),
-        output_path=args.get("output_path")),
+        output_path=args.get("output_path"),
+        speed=args.get("speed")),
     check_fn=check_tts_requirements,
     emoji="🔊",
 )