I’ve yet to watch that vide but looks like that one is implemented with openai realtime API already? What do you mean by fully implement it with n8n that’s different than that approach?
In this case, openai realtime api calls n8n as it’s backend.
The openai and twilio api are managed via code (on replit)
I would have liked to manage the web socket connections to twilio and openai realtime api from n8n.
Openai realtime api actually supports webrtc now, so it might be easier to use