Open Voice

Give your OpenClaw agent a voice. Five minutes. Two keys. That's it.

Paste your Discord bot token and your ElevenLabs API key. Your OpenClaw agent handles the rest — clones the repo, builds Whisper, configures everything, starts the bot. You just talk.

How it works: You speak → local Whisper transcribes your voice (free, runs on your machine) → your OpenClaw processes it with full context (files, memory, tools, everything) → ElevenLabs speaks the response back in your voice channel.

Like a phone call with your AI agent. Hands-free. Walk around. Cook dinner. Drive. Your agent is in your ear.

What you get:

• Private GitHub repo access (lifetime, all future updates)
• SKILL.md your agent reads and follows automatically
• Local Whisper transcription — $0 for speech-to-text
• Works with any ElevenLabs voice
• Say "stop" to interrupt mid-speech
• Runs on your hardware — your keys, your data, your privacy

Setup:

1. Paste your Discord bot token
2. Paste your ElevenLabs API key
3. Your OpenClaw agent does everything else

That's literally it. No config files. No terminal commands. Your agent reads the skill and sets itself up.

Requirements: OpenClaw running on a Mac or Linux machine, Node.js 18+, FFmpeg