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.

There are some (fantastic new models)[https://huggingface.co/spaces/hf-audio/open_asr_leaderboard] coming out all the time, this is a thin wrapper so PR a model that you prefer and we can keep adding them to the list.