Speak.
It’s text.
On-device dictation for Windows, macOS & Linux. Press a hotkey, talk, and your words land in any app, transcribed locally. No cloud. No account. No telemetry.
Reply to the team that the build is green and shipping at noon.
Built for speed, sworn to silence.
On-device inference
Your voice never leaves the machine. Local Whisper-class models, zero network round-trips.
Works in any application
A global hotkey drops text into editors, browsers, chat, email, wherever the cursor is.
Your models, your choice
Swap open ASR models (Whisper, Parakeet, Moonshine and more) to balance speed and accuracy.
Cross-platform, native
Windows, macOS on Apple Silicon, and Linux on Ubuntu / Debian.
Free and open source
Read every line, file an issue, or build it yourself. Nothing hidden.
Press your hotkey
A global shortcut starts listening from anywhere, no window to switch to.
Speak naturally
Talk as you normally would. Verbatim transcribes on-device in real time.
Your text appears
Words type straight into whatever app you’re in. Release to stop.
Your voice never leaves your device.
Questions, answered.
Is Verbatim really free?
Yes. Free and open source, no subscription, no account. The full source is on GitHub.
Does my voice leave my device?
Never. Capture, transcription, and the final text all happen locally. No cloud, no telemetry.
Why does my OS warn on first launch?
The build is unsigned. On Windows: More info → Run anyway. On macOS: right-click → Open. One time only.
Does it work offline?
Yes. Once a model is downloaded, Verbatim works with no network connection.
Can I change the hotkey or model?
Yes. Set any global hotkey, and switch between open models (Whisper, Parakeet, Moonshine, and more) in settings.
Which platforms are supported?
Windows (x64), macOS (Apple Silicon), and Linux (Ubuntu/Debian).
Get Verbatim.
All releases & checksums ↗ · Unsigned build; SmartScreen / Gatekeeper may warn on first launch (More info → Run anyway, or right-click → Open).