software · Picked by us
The offline dictation app built after breaking a finger
A free cross-platform app that runs Whisper or Parakeet speech recognition locally and pastes the result into any text field. 32k GitHub stars, signed releases for Windows, macOS and Linux.
The maker on the broken finger, the first build and the privacy choice
MIT source; free signed releases for Windows, macOS and Linux
What it does
Dictate anywhere on your machine with nothing leaving it: local speech recognition, no cloud round-trip.
Who it helps
People who write a lot and want dictation without sending audio to a server.
How it was made
- Models
- Whisper · runs locallyParakeet · alternative local model
- Tools
- Rust · Desktop app implementationLocal inference · Transcription on your machine
- CJ Pais broke a finger and wanted dictation that did not depend on a cloud service, so he built the first version in Rust.
- The app runs open speech models locally; model files download separately, and audio never leaves the machine.
- Signed releases ship for all three platforms, and the repository shows continuous fixes from user reports — including paste, Bluetooth and Wayland edge cases.
What the person did
Pais chose local processing, built and ships the app, and answers user reports directly; the repository reserves the Handy name and brand assets.
What we know
Creator-reported: The repository (32.2k stars, 3.0k forks, 860 commits on 2026-09-27), the maker’s launch post and the live site; permissions and model downloads are documented limits.
Access checked — 2026-09-28 · Opened the repository, the launch post and the live site on 2026-09-28; counts are from the 2026-09-27 check.
Can you use it?
The source is MIT; the Handy name, logo and icon are reserved by the maker. We link only and host no screenshots. The cover is an original Made With Models illustration, not a screenshot of the app.
About this project
Picked by Made With Models. Added 2026-09-28. Updated 2026-09-28.