After successfully providing provider credentials, when I try to call an assistant and ask if something, I get an error saying "Error: Exiting meeting because room was deleted" and the call ends. When I remove my preovider keys though, this error disappears and the call works fine.

Hey can you share the call ID?

We had the same issue mid week, and we're now seeing it for a high % of call regardless of no longer having the custom openAI key present. 🤔

openai changed how they return tool call, and because of this calls were failing. Now the fix is live so you can try again.

I tried again but it's still not working. The call id is 2bc39cb3-d1b8-4cea-b978-5cdf8bd223ab

Looking at your error message more specifically: 🟡 08:17:55:502 No model found to process LLM request 🟡 08:17:55:505 No Available Model For { "provider": "openai", "model": "gpt-4o-mini", "fallbackModels": \[\], "temperature": 1, "maxTokens": 250, "functions": \[\], "tools": \[\], "toolIds": \[\], "messages": \[ { "role": "system", "content": "You are an assistant." } \], "numFastTurns": 0, "semanticCachingEnabled": false, "emotionRecognitionEnabled": true } The specific issue is that you're trying to use gpt-4o-mini but: \- Either the OpenAI credentials are missing in your configuration, or \- Your OpenAI account doesn't have access to the gpt-4o-mini model The system filters available models based on credentials. In your case, you only have: \- An Azure credential for speech service \- An Azure OpenAI credential with access to only gpt-35-turbo-0125 model The solution would be to either: \- Add OpenAI credentials with access to gpt-4o-mini, or \- Change the model to use gpt-35-turbo-0125 since you have Azure OpenAI credentials for it: