LocalVoiceAI is a free, fully on-device voice transcription tool for macOS. Hold Fn+F10 in any app — Claude, Cursor, VS Code, your browser — speak your thought, release the key, and the transcribed text is pasted directly at your cursor. Everything runs locally on Apple Silicon using the Whisper model via Metal GPU acceleration. No audio ever leaves your machine.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/npateriya/LocalVoiceAI/llms.txt
Use this file to discover all available pages before exploring further.
Quickstart
Download, install, and be dictating in under five minutes.
Permissions Setup
Grant the one-time macOS privacy permissions required to run.
Configuration
Change the push-to-talk key and customize behavior.
Troubleshooting
Fix common issues like event tap failures and gibberish output.
Why LocalVoiceAI?
| LocalVoiceAI | Cloud STT (Whisper API, etc.) | |
|---|---|---|
| Cost | Free | Pay per minute |
| Privacy | Fully on-device | Audio sent to cloud |
| Latency | ~1–2s (Metal GPU) | Network dependent |
| Works offline | Yes | No |
| Setup | One-time | API key required |
How it works
Hold Fn+F10
A system-wide key tap starts recording from your microphone the moment you press the key.
Speak
PortAudio captures audio at your mic’s native sample rate (48kHz / 44.1kHz) with no resampling artifacts.
Requirements
- Apple Silicon Mac — M1 or later (Metal GPU required)
- macOS 14+
- Homebrew — for
whisper-cppandportaudiodependencies
LocalVoiceAI runs as a macOS LaunchAgent so it works consistently whether you launch it from Terminal, iTerm2, VS Code’s integrated terminal, or anywhere else — no terminal window needs to stay open.