Theo - t3.gg
Due to my hand injury, I have been doing a lot of voice-to-text lately. Despite this, my frustrations with VTT have only grown. There are so many things these tools and models get wrong. I assumed there would be some way to customize, but I've learned that you really don't have much control before the model does the transcription. All of the nice cleanup features in tools like Wispr Flow are done AFTER the text is generated. I'm so incredibly frustrated with this and I'm trying to decide if I care enough to fix it.
