This is a simple ish python implementation to connect to the new RealTime OpenAI API.
I've watched multiple videos of people struggling to deal with interrupting the speech.. in probably 5 prompts with o1 preview - this what I got.
Seems to work great on Windows, not tried linux yet.
Just put your api key in .env (check example) and run main.py, and it will ask you to select your audio input, output, and voice to use - then you can freely talk real time back and forth and interrupt.
This should be useful as a base for other applications or just to see how one may or may not do it.