Ask HN: How to replace macOS default dictation

I try to use Speech to Text(stt) models as much as I can when programming to reduce the amount of inputI have to type as I have carpal tunnel syndrome. On macOS, the default stt model is mediocre compared to the progress made by other models like Whisper.

For running a stt model locally, what models are recommended? Anybody already working on this?

Any links, repos, or advice would be greatly appreciated.