Do you know a really accurate speech-to-text model (other way round)?
I’ve always just relied on Apple on device method and that seemed to work well - but obviously only in apples ecosystem. Not sure about models.