Beyond Flow
Voice dictation that runs on your own GPU. Whisper large-v3-turbo does the work locally, a global hotkey starts it, and the text lands in whatever window you were already working in.
- Windows 11
- Windows 10
- x64
- Electron
- Python
- faster-whisper / CTranslate2
- CUDA
Built, tested and gated. Release is held until code signing is in place.
Our builds are finished and gated, but they are not yet signed with a publicly-trusted certificate. Windows would present them as an unknown publisher. We are not shipping software that teaches you to click past the warning that protects you. Releases open the day signing is in place.
We are at a keyboard all day, running several agents and several processes at once, and our hands could not keep up with our own work.
- Dictation was the obvious fix, so we paid for it.
- The licence turned out to be charged per device.
- We work across three machines, which meant three subscriptions.
- The model would run happily on the card already sitting in the machine.
Paying three times over to talk to our own computers was where we stopped. We wrote our own that evening.
Stop typing. Nothing leaves the machine.
- Verification gate 15 of 15 passing
- No dependency on any other Beyond application
Documentation
Installation, requirements, configuration and licensing for Beyond Flow.
Release history
Every release, what changed in it, and the hash of exactly what shipped.