IndexTTS2 Voice Cloning API

Service: IndexTTS2 Zero-shot Voice Cloning with Queue Control
Concurrency Limit: 1
Queue Management: Enabled
Response Format: JSON with download URL
CORS: Enabled

API Endpoints:
  POST /tts              - Text to Speech (returns JSON)
  GET  /download/<file>   - Download generated audio file
  GET  /health            - Health check
  GET  /queue/status      - Queue status monitoring
  GET  /docs              - Swagger API documentation
  GET  /redoc             - ReDoc API documentation

Request Parameters (POST /tts):
  Required:
    - text                 : Text to synthesize (1-5000 characters)
    - reference_audio_url  : URL of reference audio file

  Optional (Emotion Control, 0-100 for each):
    - emotion_happy       : Happy emotion (default: 0)
    - emotion_angry       : Angry emotion (default: 0)
    - emotion_sad         : Sad emotion (default: 0)
    - emotion_fear        : Fear emotion (default: 0)
    - emotion_disgust     : Disgust emotion (default: 0)
    - emotion_melancholy  : Melancholy emotion (default: 0)
    - emotion_surprise    : Surprise emotion (default: 0)
    - emotion_calm        : Calm emotion (default: 0)
    - use_random          : Enable randomness (default: false)

Response Format (POST /tts):
  Success Response:
    {
      "success": true,
      "message": "TTS generation completed",
      "data": {
        "request_id": "indextts_123_20260316_120000",
        "download_url": "http://YOUR_SERVER:8001/download/indextts_123_20260316_120000.wav",
        "filename": "indextts_123_20260316_120000.wav",
        "processing_time": 5.35,
        "text_preview": "Hello, this is a...",
        "reference_url": "https://example.com/reference.wav",
        "emotions": {
          "happy": 80,
          "calm": 20
        }
      }
    }

Technical Specifications:
  - Model Version        : IndexTTS2
  - Audio Format         : WAV
  - Concurrency Mode     : Async Queue
  - Concurrency Limit    : 1
  - Zero-shot Cloning    : Yes
  - Emotion Control      : 8-dimensional emotion vector
  - Response Format      : JSON
  - Audio Storage        : Temporary files with download URLs
  - Audio Source         : HTTP/HTTPS URL
  - Audio Caching        : Yes (MD5 hash based)
  - CORS                 : Enabled

Important Notes:
  1. API returns JSON response with download URL
  2. Audio files are stored temporarily and available for download
  3. Download URL format: https://index-tts.dzai.cc/download/<filename>
  4. Reference audio is downloaded from URL and cached
  5. Concurrency controlled by 'concurrency_number' parameter
  6. CORS enabled for cross-origin requests

Powered by IndexTTS2 | Version 2.0.0 (Queue Control + JSON + CORS)