fix backend TTS provider compatibility

backend/models.py

@@ -26,6 +26,7 @@ class AssistantConfig(BaseModel):
     # 模型/音色选项
     model: str = ""  # LLM
     asr: str = ""    # STT
+    tts_model: str = ""
     voice: str = ""  # TTS 音色
     stt_language: str = ""
     tts_speed: float = 1.0

backend/services/config_resolver.py

@@ -58,6 +58,7 @@ async def resolve_runtime_config(
         # 模型/音色:凭证的模型ID优先
         model=(llm.model_id if llm else ""),
         asr=(stt.model_id if stt else ""),
+        tts_model=(tts.model_id if tts else ""),
         voice=(tts.voice if tts else ""),
         stt_language=(stt.language if stt else ""),
         tts_speed=(tts.speed if tts else 1.0),

backend/services/credential_tester.py

@@ -84,6 +84,7 @@ async def test_openai_credential(
                         "model": config.model_id,
                         "input": "测试",
                         "voice": config.voice,
+                        "response_format": "pcm",
                         "speed": config.speed,
                     },
                 )

backend/services/pipecat/service_factory.py

@@ -10,7 +10,7 @@ from models import AssistantConfig
 
 from pipecat.services.openai.llm import OpenAILLMService
 from pipecat.services.openai.stt import OpenAISTTService
-from pipecat.services.openai.tts import OpenAITTSService
+from pipecat.services.openai.tts import VALID_VOICES, OpenAITTSService
 from pipecat.transcriptions.language import Language
 
 
@@ -48,12 +48,18 @@ def create_llm(cfg: AssistantConfig):
 
 def create_tts(cfg: AssistantConfig):
     """CosyVoice 等,走 OpenAI 兼容的 /v1/audio/speech。"""
+    voice = cfg.voice or config.TTS_VOICE
+    # Pipecat 默认只接受 OpenAI 官方音色。OpenAI 兼容服务常使用自定义 voice id,
+    # 注册为原样映射后仍由 OpenAI SDK 按字符串透传给供应商。
+    VALID_VOICES.setdefault(voice, voice)
     return OpenAITTSService(
         api_key=cfg.tts_api_key or config.TTS_API_KEY,
         base_url=cfg.tts_base_url or config.TTS_BASE_URL,
-        model=config.TTS_MODEL,
+        settings=OpenAITTSService.Settings(
-        voice=cfg.voice or config.TTS_VOICE,
+            model=cfg.tts_model or config.TTS_MODEL,
-        speed=cfg.tts_speed,
+            voice=voice,
+            speed=cfg.tts_speed,
+        ),
     )
 
 
@@ -61,6 +67,7 @@ def create_services(cfg: AssistantConfig):
     logger.info(
         f"创建服务: stt={cfg.asr or config.STT_MODEL} "
         f"llm={cfg.model or config.LLM_MODEL} "
-        f"tts={cfg.voice or config.TTS_VOICE}"
+        f"tts={cfg.tts_model or config.TTS_MODEL} "
+        f"voice={cfg.voice or config.TTS_VOICE}"
     )
     return create_stt(cfg), create_llm(cfg), create_tts(cfg)