This is a simple tool that is designed to make it feasible to dictate text using the whisper speech to text model from open AI. Whisper is not designed for dictation, so it requires quite a bit of fiddling to make it work nicely.Whisper is only really designed for transcription and translation.
Run the script and then start talking. The text should appear in the application and emulate your keyboard.Note that there is significant amounts of latency due to the architectureof the Whisper Machine Learning Model.
- Reduce Latency
- Add cli arguments
- Add a better UX