https://vapi.ai/ logo
Join Discord
Powered by
High latency
# support
l
I'm experiencing a high latency of 29 seconds while using our AI Assistant, which is unusual compared to previous performance when it is around 2 seconds. This issue began recently. Was working fine yesterday. Could you please investigate and resolve this as soon as possible? Call ID: 383d1668-e4a1-45a9-b7b9-bd992cd13626
v
message has been deleted
@Lajiao I'm sorry to hear that you're experiencing high latency with the AI Assistant. To help you effectively, I'll need to gather a bit more information. Could you please confirm if there have been any changes to your network setup or if you've noticed any other unusual activity that might correlate with the issue you're facing? Additionally, it would be helpful to know if this latency is occurring consistently or if it's intermittent. In the meantime, I'll initiate a preliminary check on the system status using the Call ID you've provided to see if there are any immediate issues that can be identified. Please hold on for a moment while I perform this check.
You can continue this conversation by mentioning me in the message.
l
@Vapi Nothing has been changed since yesterday.
v
@Lajiao Thank you for confirming that there have been no changes on your end. I will proceed with checking the system status using the Call ID you provided to identify any issues that could be causing the high latency. Please bear with me for a moment while I look into this for you.
s
Can you provide some more call_ids? @Lajiao
Also, from when this issue started occuring?
l
I just made two more calls: 921937d0-ff5d-4170-9b47-3e8d693db77e 2add80be-64f9-448f-ba28-66d6c1358c3b We haven't put this into production yet, so we test from time to time. I just noticed this issue right before I opened this ticket. I made an outbound call with an English-speaking assistant, and it was working fine. Maybe it is because I am using Eleven Turbo 2.5 with the Swedish language. I just tried changing it to the Multi-Language setting, and it worked. Then I switched it back to Eleven Turbo 2.5, and now it is working again. So, I am fairly positive it has to do with Eleven Labs.
s
Yeah. I see let me check.
Can you point me to the exact timestamp cause I don't know swedish language
l
I made some more tests now and still got problem. It seems like it's the biggest problem with Elven Turbo 2.5 but also the multi language are slow. Maybe ElevenLabs have some issues now.
s
Give it a try to Cartesia heard lot of good things about them
l
Do you want a specific time stamp in a call? Which call id in that cass?
s
921937d0-ff5d-4170-9b47-3e8d693db77e
2add80be-64f9-448f-ba28-66d6c1358c3b
l
I have been trying different once but I will give it a try. Thanks for suggestion.
@Sahil I was trying Cartesia and it was 6 seconds silence there as well.
@Sahil I did some more research, and this is the only assistant using the GPT-4o Mini Cluster. I tried different voices from various services, and it doesn't seem like the voices are the problem. When I switched to the GPT 4o Cluster, the response was immediately faster. However, there is still more delay than expected, so I am beginning to think the problem might be on OpenAI's side. I made at least 15 calls during these tests.
2 Views
PreviousNext
Powered by