No errors or status codes, no. And looking at the Vapi dashboard's Logs entries shows nothing unusual that I can see (websockets connections to OpenAI are working, etc). It's just like the agent isn't even there. Here is an example Call ID of one happening directly in the Vapi console using our primary Assistant: 019e4759-42b8-7994-bcde-eaf4702cbe21 ...and here is one from our full setup (via Twilio, etc): 019e471e-1301-7220-85d8-0d8b4f95badd Also, if I create a new Assistant using the GPT Realtime model, it does not work even with zero custom config. So I agree it's something more basic - and I imagine you should be able to repro? Unless there is something broken at our account level...