Looks like it solves the speech to text piece but not text to speech? Neat tool either way.

I am looking for any examples of people fully working with their coding agent CLI via voice alone.