Hey , this usually happens when Vapi hits a brief hang on WS/WebRTC and the session doesn’t recover, so it wrongly logs it as “user terminated.” I’ve fixed the same issue for Soltan and Francis both had random call drops caused by stalled responses and WebRTC heartbeat issues, not user/network problems. Quick questions: Are you using tools/functions during the call? Does it happen after a specific step or randomly? Any custom backend in the flow? I can help you pinpoint and fix it quickly 👍

Hi Amos, thanks for the quick reply. I’m currently out so I won’t be able to try much now. But basically they’re pretty simple agents. No tooling or functions. Just straight calling, talking with the agent and finishing.

Got it ,that actually rules out your setup. If nothing changed on your side and it suddenly started today across both WS + WebRTC, it’s very likely a Vapi-side instability or timeout regression, not your agents. The random timing during calls also points to session drops rather than flow issues. I saw this exact pattern recently, same behavior, no changes, and it cleared once Vapi stabilized. For now, I’d suggest: Check Vapi status/logs for today Grab 1–2 failed call IDs for support Once you’re back, I can help you trace it properly 👍