After months of vibe-coding, I found myself not committing to any speech to text app.
Mostly because I didn't trust them:
I wanted something that ran locally, was open-source, and wasn't trying to squeeze me for a subscription for a trivial wrapper around a plethora of models.