@Vapi Incident @Shubham Bajaj All incoming calls are having issues with lost messages during transcription and communication, which leads to multiple repeats of same messages and uresponsiveness. All outgoing calls are getting constantly disconnected after a second. Examples of incoming calls: b29acde0-1f85-47ae-841c-9883d3734ec8 529bf65d-3e0f-40b6-8a74-eb66b73f96cd b43635cb-58b8-4380-b347-00f1abf3635a Examples of outbound calls 48b1bd5c-89c2-458d-a894-f1d6fbfecb86 b2392830-f959-43c4-abbf-548d4307a56c 9b1e3aad-7526-47a2-9d72-b57c15456cd7 Twillio is configured to do a post webhook for voice to : https://api.vapi.ai/twilio/inbound_call

Definatelly, however pointing same number to 11labs call endpoint does not result in any issues

It doesn’t seem like a widespread outage but for sanity check I will check with the team again about it.

sure, i am surprised if i would be the only one experiencing this, as every outbound call is getting dropped in a matter of 5 seconds, and multiple inbound numbers suffer from lost messages which results in constant repetition. again, pointing same number to 11 labs webhooks results in smooth experience, so it leads me to believe that it is not twillio isolated issue but twillio -> vapi issue.

I tested earlier and responses were better. But not perfect from the 'drift' issue. (Australia)

was able to successfully migrate all numbers to 11 labs and retell ai without of a hint of issue. So seems like the solutions currently is to move to another service is you are experiencing simmilar issues

My clients account yes, my vapi account no, and both twilio... so 🤷🏻‍♂️

@Panda @~Stu~ CoreAspect.Agency There was an issue from Twilio's side which resulted in a mix-up of audio and LLM outputs, creating an issue for you. I'll suggest you try making 2-3 test calls and let me know how it goes. I'm pretty sure the issue has been resolved by Twilio's team, so you won't be experiencing it again.

My last call was perfect. So I'm hoping this is all good. One thing I'm unsure of tho, and maybe you can help me reason with( but i understandits not your system but may understand more of the underlyingarchitecture). Why was it OK on one vapi account and not the other (same backend on my part). Is it because twilio use different servers clusters for different phone number segments? I'm just trying to get my head round why

That issue was for UK clusters only (or I'll say UK regions), and it wasn't affecting other regions, which is why you had a issue with particularly with few numbers only.