Note: New documentation is available at https://doc.livetalking.ai

Welcome to livetalking documentation!

Note

This project is under active development.

Contents

• 0. Features
• 1. Installation
  • 1.1 Install dependency
• 2. Quick Start
• 3. Usage
  • 3.1 Digital Human Model
    • 3.1.1 Model: wav2lip
      • Use your own avatar
    • 3.1.2 Model: musetalk
      • Use your own avatar
    • 3.1.3 Model: ER-Nerf
      • 3.1.3.1 Audio feature: hubert
      • 3.1.3.2 Set head background image
      • 3.1.3.3 Full-body video overlay
      • Use your own avatar
    • 3.1.4 Model: Ultralight-Digital-Human
  • 3.2 Transport Mode
    • 3.2.1 WebRTC P2P
    • 3.2.2 WebRTC push to SRS
    • 3.2.3 RTMP push
  • 3.3 TTS Model
    • 3.3.1 gpt-sovits
    • 3.3.2 fish-speech
    • 3.3.3 cosyvoice
    • 3.3.4 Tencent Cloud TTS
    • 3.3.5 Doubao (Volcengine) TTS
    • 3.3.6 Alibaba Qwen TTS
    • 3.3.7 XTTS
  • 3.4 Action Choreography
  • 3.5 LLM Dialogue
  • 3.6 Multi-Concurrency
  • 3.7 Audio Input
  • 3.8 Virtual Camera Output
• 4. Docker Run
• 5. API Interface
  • 5.1 Get Video
    • 5.1.1 Request Parameters
    • 5.1.2 Response Data
  • 5.2 Send Text
    • 5.2.1 Request Parameters
  • 5.3 Switch Playback Video
    • 5.3.1 Request Parameters
  • 5.4 Recording
    • 5.4.1 Request Parameters
  • 5.5 Interrupt Digital Human Speech
    • 5.5.1 Request Parameters
  • Third-Party Software Integration Process
• 6. Performance Optimization
  • 6.1 Image Caching
• 7. Paid Services
  • 7.1 HD Version of wav2lip
  • 7.2 Feature Optimization