FlowVoice
Installation
SKILL.md
Flow Voice — Voice Cloning for OpenClaw
Clone any voice from a 3–30 second audio sample and generate speech from text. Powered by LuxTTS — 150x realtime, runs locally, fits in 1GB VRAM, works on CPU and Apple Silicon MPS. No API key, no cloud, no cost.
Output directory: ~/clawd/output/voice/
Commands
| What you say | What it does |
|---|---|
| "clone this voice [audio file]" | Encode a voice profile from a sample |
| "speak as [name]: [text]" | Generate speech using a saved voice profile |
| "add voiceover to [video]: [text]" | Generate speech + bake into video with ffmpeg |
| "list voices" | Show saved voice profiles |
| "clone voice from URL [url]" | Download audio from URL, then clone |