@epiphytic/openclaw-discord-voice

OpenClaw plugin for real-time Discord voice assistant with fully local AI processing.

Features

• Voice conversations — join Discord voice channels and interact via speech
• Local STT — whisper.cpp with GPU acceleration
• Local LLM — any OpenAI-compatible model (Qwen, GLM, etc.)
• Local TTS — Kokoro-82M for natural speech synthesis
• Tool calling — weather, time, web search
• Escalation — complex requests route to the main OpenClaw agent seamlessly
• Text-to-voice bridge — text channel messages read aloud in voice
• Speaker identification — optional Resemblyzer-based speaker recognition

Requirements

• Python 3.11+ with py-cord, whisper.cpp, Kokoro TTS
• Local LLM server (vLLM, llama.cpp, etc.) with OpenAI-compatible API
• Discord bot with Voice, Message Content, and Server Members intents

Install

openclaw plugins install @epiphytic/openclaw-discord-voice

Configuration

{
  plugins: {
    entries: {
      "discord-voice": {
        enabled: true,
        config: {
          botToken: "your-discord-bot-token",
          guildIds: [123456789],
          botName: "Assistant",
          llmUrl: "http://localhost:8000/v1/chat/completions",
          whisperUrl: "http://localhost:8001/inference",
          kokoroUrl: "http://localhost:8002/v1/audio/speech"
        }
      }
    }
  }
}

See the repository for full configuration options and setup guide.

License

MIT