更新TTS相关配置，调整语音ID格式，优化工作流参数，确保一致性和可读性。

api/routers/tts.py

@@ -23,7 +23,7 @@ async def tts_synthesize(
     Convert text to speech audio.
     
     - **text**: Text to synthesize
-    - **voice_id**: Voice ID (e.g., 'zh-CN-YunjianNeural', 'en-US-AriaNeural')
+    - **voice_id**: Voice ID (e.g., '[Chinese] zh-CN Yunjian', '[English] en-US Aria')
     
     Returns path to generated audio file and duration.
     """

api/schemas/tts.py

@@ -8,13 +8,13 @@ from pydantic import BaseModel, Field
 class TTSSynthesizeRequest(BaseModel):
     """TTS synthesis request"""
     text: str = Field(..., description="Text to synthesize")
-    voice_id: str = Field("zh-CN-YunjianNeural", description="Voice ID")
+    voice_id: str = Field("[Chinese] zh-CN Yunjian", description="Voice ID")
     
     class Config:
         json_schema_extra = {
             "example": {
                 "text": "Hello, welcome to ReelForge!",
-                "voice_id": "zh-CN-YunjianNeural"
+                "voice_id": "[Chinese] zh-CN Yunjian"
             }
         }
 

api/schemas/video.py

@@ -23,7 +23,7 @@ class VideoGenerateRequest(BaseModel):
     
     # === Basic Config ===
     n_scenes: int = Field(5, ge=1, le=20, description="Number of scenes (generate mode only)")
-    voice_id: str = Field("zh-CN-YunjianNeural", description="TTS voice ID")
+    voice_id: str = Field("[Chinese] zh-CN Yunjian", description="TTS voice ID")
     
     # === LLM Parameters ===
     min_narration_words: int = Field(5, ge=1, le=100, description="Min narration words")
@@ -57,7 +57,7 @@ class VideoGenerateRequest(BaseModel):
                 "text": "Atomic Habits teaches us that small changes compound over time to produce remarkable results.",
                 "mode": "generate",
                 "n_scenes": 5,
-                "voice_id": "zh-CN-YunjianNeural",
+                "voice_id": "[Chinese] zh-CN Yunjian",
                 "title": "The Power of Atomic Habits"
             }
         }