Hi, I have been working with VAPI over the past 3 months developing a new AI Assistant app. Today the assistant is sounding like they are in slow motion and I see extreme latency issues. It is a system problem today or should I contact support to talk about this issue specifically?

This doesn’t look like a VAPI system issue. Sudden slow voice and high latency usually mean two databases or services aren’t properly automated or linked, so the assistant is waiting on delayed data. The fix is to set up an automatic trigger or real-time sync so data flows instantly. I can help you get that dialed in quickly. What platform or database is VAPI connected to right now?

Hi any help is greatly appreciated, thank you. I am setup using Lovable/Supabase and deploying to Netlify. Supabase edge functions are used for the tool calls from VAPI.

Which models are you using. Also have you checked call logs to see latency summaries of each call to see where the latency is coming from specifically

checking the latency summary was a good catch. looking at that it looks like my endpointing is taking 1,000-2,000ms. I will check to see if that smart endpointing setting is still on. but, would this cause the slow, slured, slomotion speach?

Thanks for sharing that setup explains the latency. With Supabase edge functions and VAPI tool calls, this usually happens when data isn’t synced in real time and the assistant is waiting on delayed responses. I help fix this by tightening the edge functions and adding proper triggers so everything updates instantly. Happy to help you with this feel free to message me and we can talk details. @Pancho

Vapi's endpointing is extremely long. I honestly just completely turned off livekit and on start/stop speaking plan set the agent to wait only 0.2 seconds. Atleast with english i havent had any issues with the assistant talking over or interupting, just make sure to set start speaking to after caller has said like 2-3 words.