Yeah, I can help with that, you’re spot on that those org and assistant server URLs add real latency because VAPI queues them before the model runs.
The solution is to remove the server URLs from both Organization and Assistant settings and keep webhooks only on the tools that actually need them.
That way, only tool calls trigger requests and your conversation flow stays fast and responsive.
Just to confirm, are you currently using those global webhooks for anything critical like transcripts or call logging?
@Argonautas