discord-voice

What This Skill Does

The discord-voice skill transforms your OpenClaw agent into an interactive, real-time voice companion for your Discord server. It enables the bot to join voice channels, listen to incoming audio, transcribe speech using advanced engines like Whisper or Deepgram, process that text through the Claude AI core, and return natural-sounding speech via TTS providers like ElevenLabs or Kokoro. This skill bridges the gap between text-based AI interaction and fluid, human-like voice communication within your existing community infrastructure.

Installation

1. System Dependencies: Ensure ffmpeg, build-essential, and python3 are installed on your host machine.
2. ClawdHub Install: Run clawdhub install openclaw/skills/skills/avatarneil/discord-voice.
3. Manual Setup: Clone the repository into ~/.clawdbot/extensions/discord-voice, run npm install, and verify your binary dependencies are compiled.
4. Configuration: Add the plugin to your clawdbot.json configuration file, ensuring you provide the necessary API keys for your preferred STT/TTS providers.
5. Discord Permissions: Update your bot's application settings to enable 'Connect', 'Speak', and 'Use Voice Activity' permissions.

Use Cases

• Voice-Activated Assistance: Allow server members to ask questions or get help without leaving the voice channel.
• Language Learning: Use the bot as a conversational partner to practice new languages in real-time.
• Meeting Transcription & Summarization: Have the bot listen to voice channel discussions and provide automated meeting minutes.
• Interactive Entertainment: Integrate the agent into roleplay or gaming sessions for immersive character interaction.

Example Prompts

1. "@Clawbot, please join the Voice Lounge and let's discuss the latest open-source project updates."
2. "@Clawbot, start listening to this channel and summarize the main points everyone makes during our meeting."
3. "@Clawbot, leave the voice channel; we are finished with our conversation for now."

Tips & Limitations

• Performance: Use local models if you have high-performance hardware, or cloud APIs for better speed and lower CPU usage.
• Barge-in: The barge-in feature is sensitive to background noise; ensure your microphone threshold is calibrated correctly in your config.
• Latency: Use Deepgram WebSocket support for the lowest possible latency in multi-turn conversations.
• Security: Be mindful of which channels the bot has access to, as it will record all audio within those channels when active.