Use cases I readily see for the possibility of aborting the speech processing:
some kind of cancellation word or sound: "all that I just said is rubbish, please discard"
having a "drowse" type word where everything after is discarded, not processed or shown as subtitles, so you can both pause Talon and continue speaking to others in the same breath
In order to do things like:
it would be useful if there were a function that could be used to clobber Talon's audio buffer.
We can simulate this behaviour today by cycling the mic, but that messes up the mic history
Obv entirely up to you what the api should look like, but something as simple as
speech.clear_audio_queue()
or something like that would work for me