OpenAI WebRTC Audio Session, now with document context
Simon Willison updated his OpenAI WebRTC audio tool to support GPT-Realtime-2 and pasted document context for voice conversations.
“Paste a document here before starting the session and the model will be able to discuss it with you”
Simon Willison revived his browser-based OpenAI WebRTC audio playground, adding support for the new GPT-Realtime-2 voice model and the ability to paste document context to have spoken conversations about it. It's a useful demonstration of OpenAI's realtime voice API but is a personal hobby tool rather than a major industry announcement.