As of about 30 minutes ago, I no longer am getting any voice responses via the Realtime API, even when using the default client/prompt via OpenAI-Realtime-Console github project and a fresh API key. Has anyone else seen this? Does anyone have any debugging tips? I can see the Whisper transcriptions of my own messages but the model itself does not respond in any way.
I figured it out - the issue was with rate limits. It would be nice if the demo app caught this scenario and warned the user. In any case it’s not a bug, please ignore this thread!