This is my personal version of the (so far nonexistent) Onyx Systems core microservices. This is the original project and might change over time, but is completely self-hostable. Still very much pre-production though.
It seems I was confused about what TE was, it doesn't correct your words, just formats your text. This functionality has been added but I'm starting a new issue to deal with adding autocorrect
The STT model on the speech server is pretty good, but not good enough. I want to see if I can significantly enhance it with TE silero models.