As I wait for my parts to print, I'm writing some code for the voice assistant. Just did a little outline of how I think it will go, but I haven't tested anything yet:
https://github.com/potentiallyunwanted/SCOUT-voice-assistant

right now the idea is that it's always checking audio -> waits for it's named to be called "scout" (but can be anything)
takes human response -> sends to openai
openai -> elevenlabs for awsome british voice -> plays response

adding a conferencing that will attach via usb for audio and mic
hopefully it all works out.