Press a hotkey, speak naturally, and PressType transcribes your words directly into any app. Local Whisper models, zero cloud dependency, built for developers.
One hotkey triggers recording. Speak naturally. Release, and your words appear in the active app — VS Code, Slack, Terminal, anything. No windows, no friction.
Whisper runs entirely on-device via whisper.cpp. Audio
is processed in memory, never written to disk, never sent to a
server.
Add developer terms to your vocabulary. Spelling corrections and keyterm boosting ensure technical accuracy out of the box.
VS Code uses keyboard injection. Slack uses clipboard paste. PressType auto-detects the active app and applies the right profile — or you set your own.
Download Whisper Large for accuracy, Turbo for speed, or Belle for Chinese. Connect cloud APIs when you need them. Switch in one click.
Track presses, words, latency, and per-app breakdown. A GitHub-style heatmap shows your active hours at a glance. Filter by time range.
A warm, focused dashboard that tells you how you're using your voice — without getting in the way.
Run entirely local with Whisper, or connect cloud APIs for speed. Switch models in one click.
Free, open source, runs entirely on your machine. Available for macOS with Apple Silicon.