currently my latency is too much, it's like 5 seconds or something, trying to get it down... but what is a good way to troubleshoot why the latency is this high?

If you go into vapi dashobard at the top of your assistant you should see latency details. It will show you exactly what is causing the latency.

yes but that's only an estimate right?

Mouse over any of the colors and it'll show what it is for https://cdn.discordapp.com/attachments/1366412051207032942/1366413705679470642/Screenshot_2025-04-28_at_8.58.39_AM.png?ex=6810db6e&is=680f89ee&hm=f728aeee800820d7f27c521dc171076d6c8da2d19af75bce9f51a78b71c042f8&

This is innacurate

It is not 100% accurate but it can give you sense of what is the bottleneck.

for me it say's 900ms but it' smuch more then that... e.g. when I measure from when I stop talking, until agent responds... it's like 5 seconds

This doenst take in mind any network latency

I understand there is also a delay of Twilio...

Okay then it is a bigger problem to investigate.

but I just want to figure out what is causing the delay

Where are you physically? 🙂

Netherlands

Yeah, so that's additional 300-400ms

but weirdly enough... if I but my voice to a Dutch (netherlands) voice... then latecy is like 2-3 seconds which is acceptable to me

You can see in call logs how long LLM takes I believe, openAI is quite janky lately, response times deviate immensely request by request

ah maybe that explains why sometimes it's fast, while other times it's slow 🙂

ahhhhh

yeah, INFO tags

thank you for shoign that

Your endpointing is quite long