VocaHQ/whisperkit-coreml
Automatic Speech Recognition • Updated • 42
Fine-Tuning ASR
Private voice dictation that runs on your own devices.
Voca apps turn speech into text where you're typing. Speech-to-text runs on the device by default, or on an optional self-hosted VocaGateway you run yourself, and never on a cloud speech service.
| App | Platform |
|---|---|
| VocaPhone | iPhone and Android keyboard (vocaphone.vocahq.com) |
| VocaMac | macOS |
| VocaLinux | Linux |
| VocaWin | Windows (beta) |
| VocaGateway | Optional self-hosted transcription server |
We publish on-device speech models here when an upstream model has no build small enough for a phone. Every model is pinned by SHA-256 in the apps, credited to its original authors under its original licence, and ships with the script that reproduces it.
vocahq.com · GitHub · X @vocahq · Discord
Voca apps are open source under AGPL-3.0.