100% on-device
Speech recognition and AI cleanup run on Apple Silicon. Your voice is never uploaded — there's no server to upload it to.
Hold ⌘, speak, release. Polished text appears in whatever app you're using — transcribed and cleaned up entirely on your Mac's Neural Engine.
Speech recognition and AI cleanup run on Apple Silicon. Your voice is never uploaded — there's no server to upload it to.
Nothing to sign up for and nothing to sync. Works on a plane, in a SCIF, or anywhere your Wi-Fi doesn't.
Cloud dictation runs $8–15 every month. Lousa is a one-time purchase — because your own computer does the work.
How it works
Every step happens between your microphone and your cursor. Nothing crosses the boundary.
Hold the hotkey and talk naturally.
On-device speech recognition on the Neural Engine.
A local AI model fixes punctuation, fillers, and tone.
Polished text lands at your cursor, in any app.
We can't see your data. There's no server, no telemetry on your speech, no transcript history in anyone's cloud.
Features
State-of-the-art on-device transcription, fast enough to keep up with your voice.
Fillers removed, punctuation fixed, sentences tightened — before the text ever lands.
Teach it your names, jargon, and acronyms once. They come out spelled right, every time.
Casual in Slack, precise in your editor. Tone adapts to the app receiving the text.
Want the transcript untouched? Flip one switch and get exactly what you said.
Picks the right models for your machine's memory — from MacBook Air to Mac Studio.
The difference
| Lousa | Cloud dictation services | |
|---|---|---|
| Voice processed on your Mac | Yes | No — sent to servers |
| Works offline | Yes | No |
| Account required | None | Sign-up + login |
| Pricing | One-time purchase | $8–15 / month |