Free, local voice dictation for Windows and Linux.
Hold a key. Speak. Release.
Hold any configurable key to record. Release to transcribe and paste instantly into whatever app is focused.
Runs OpenAI Whisper entirely on your own GPU or CPU. Your audio never leaves your machine.
Optional Claude Haiku pass removes filler words, fixes punctuation, and preserves your language and tone.
Offload to a powerful machine on your network. Works over LAN, Tailscale, or a reverse proxy with HTTPS.
Single .exe installer for Windows. One setup script handles everything on Ubuntu, Debian, and Zorin OS.
MIT licensed. No subscription, no account, no telemetry. Download, run, done.
1. Install Python 3.10+ — tick "Add to PATH" 2. Download and run Murmur-Setup-vX.X.exe 3. Add your Anthropic API key to .env (optional — AI cleanup only) 4. Hold F9 anywhere and start dictating
git clone https://github.com/R3PC0N/murmur.git cd murmur bash setup_linux.sh ./murmur.sh
The setup script installs all system packages, creates a virtual environment, and enables the AppIndicator extension for the system tray. Log out once after setup for the global hotkey to take effect.