Hold Right Option. Speak.

Audio and transcripts never leave your Mac.

The one-time ~2.3 GB model download is the only network request Momito ever makes. Transcription runs on Apple silicon via MLX, in memory, under your login — there is no cloud component, no account, and no way to send a recording anywhere.

Your history is stored locally in your home folder with owner-only permissions, and the dashboard is served to this machine only (loopback, 127.0.0.1) behind a per-launch token.

How it works

• Hold to talk

  Hold Right Option, say something, release. The transcript is typed into whatever app has focus, with punctuation and capitalization handled by the model.

• Esc discards

  Press Esc while still holding to throw the recording away: nothing is transcribed, nothing is pasted.

• Paste last transcription

  The menu keeps the last thing it typed and can put it back on the clipboard and into the focused field, for when your click landed in the wrong window.

• Menubar state

  A small schnauzer lives in the menu bar with a badge for what it is doing right now: loading, ready, listening, transcribing.

• Local dashboard

  History, stats, and replacements — spoken-form to written-form dictionary rules — in a native window, served only to your machine.

Install

1. Download the DMG from the latest release.
2. Drag Momito to Applications.
3. First open: right-click the app and choose Open. The app is unsigned, so macOS shows a one-time confirmation before it runs.
4. Grant Accessibility. System Settings → Privacy & Security → Accessibility → enable Momito. Momito notices the moment you flip the switch.
5. Speak. The model downloads once (~2.3 GB) and is then cached on your Mac — that download is the only network request Momito ever makes.
6. Microphone: macOS prompts on your first dictation. Grant once.

Updating