Posted latency on my dashboard overnight increased from 550ms to 665ms even though I changed no provider settings. I'm curious why the change... Did latency truly increase or is this just an updated representation of a more accurate measurement? My chosen stack is deepgram, gpt40 mini, 11 flash v2, web

Several factors could cause your latency increase, but network issues are a common culprit. According to Discord discussions, Deepgram's endpointing settings can significantly impact perceived latency. Try increasing the

waitSecond

parameter to 0.8 in your configuration:

javascript
// Adjust Deepgram endpointing for better latency
{
  "waitSecond": 0.8  // Default might be too aggressive at 0.1
}

This helps prevent premature endpointing that can contribute to latency issues. Sources: - [Discord discussion about Deepgram endpointing](https://discord.com/channels/1211482211119796234/1336360452208332831) - [Discord discussion about latency factors](https://discord.com/channels/1211482211119796234/1332286268805087272)

I wasn't talking about perceived latency, I was speaking about the posted latency that's listed at the top of my vapi dashboard

Hey! To help track down this issue, could you share: - The assistant ID This would really help us figure out what went wrong!