Idea: transcribe audio file

[rock3r]

Hi, love Handy. Was wondering whether we could use all its pipeline to transcribe audio from a file instead of from the microphone? Saves me from downloading/setting up the same models to do that...

[cjpais]

Yes it will come eventually, just a lot of other bug fixes coming first. Eventually if people annoy me enough about this I will just implement it

Most of the problem is in UX in the app as this feature was not really thought of, and we need to think how it fits in with everything else. And constrain it so people don't ask a million questions or new features for it. That's the main problem.

Ideally everything just works, but I think model selection, language selection, translation are all things to consider how to make them accessible in this UI, otherwise it will just feel bolted on without real thought or care. And I guess I care enough to try and work out a better solution which takes a lot of time and effort. If people have ideas they can discuss

[rock3r]

I absolutely appreciate the UX concerns here; perhaps a CLI is a good start, perhaps we want a place you can drag & drop a file to that spits out a text file or copies to the clipboard. No rush on this, of course.