I current use STT (MacWhisper) and just hold down a hotkey to speak into any application. While I’d hope that this option would provide better SST for technical things, maybe even using your code/context to steer the transcription, I think I’ll stick with my system-wide SST support that is way easier to use.

Current: hold button, talk, release button, hit enter

Voice mode: type /voice, wait for it to connect, talk, ??? Just detect when I’ve finished? Or hit a key?

Nah, I prefer being able to edit what it thinks it heard and pause for however long I want while talking without it thinking I’m done