Dictator TTS is a real-time voice dictation overlay for Linux. Offline, lightweight, and blazing fast — your voice goes directly where your cursor is.
Speech recognition runs entirely on your machine. Nothing is ever sent to any server.
Powered by OpenAI Whisper running locally. Voice to text in under a second on modern hardware.
Works with any app — browsers, terminals, editors, chat. If it takes keyboard input, Dictator works there.
A compact draggable mic button floats above all your windows — always within reach.
Say "comma", "period", or "new line" and Dictator inserts the right character automatically.
Choose your Whisper model size, microphone, language, and push-to-talk key from the settings panel.
Whisper supports 90+ languages. Switch on the fly with no restarts needed.
Pure Python, minimal dependencies. No background daemons — just run it when you need it.
Requires Python 3.10+ and FFmpeg — available in any Linux distribution's package manager.
Grab the latest release archive and extract it anywhere on your system.
Run pip install -r requirements.txt to pull in Whisper and supporting libraries.
Run python3 dictator.py, click the mic, speak — words appear wherever your cursor is.
Free to download and use. Runs entirely offline on your Linux machine.
⬇ Download for LinuxRequires Python 3.10+ · Linux (X11) · FFmpeg