I’m using Replit, integrated with Twilio and OpenAI. And have problem that seems it is about Replit.
About the mentioned critical problem, I’ve asked to Twilio Support and got an answer that means (I think) the reason of the mentioned problem may be Replit.
Firstly, let me write about my question that I’ve asked to Twilio: (shortly) … During the call, our system takes a very long time to respond to the client. …
And here is the answer of Twilio Support Team: "I looked in all the calls and see that whenever client says something, Twilio’s speech recognition system, captures it and sends this results to your backend server at https://AI-twiliosystem-call.[replitUserName].repl.co/webhook and we get HTTP response (twiml response) after 5 seconds. This is very delay here. Usually HTTP request-response are done within the same seconds but your backend server is taking too much of time in sending HTTP response. I am attaching one screenshot of one of the call SID. You will find the similar delay in all the calls.
This answer made me think that the source of the problem was “replit”.
Before asking my developers to change our server to another provider, I wanted to post the issue here. May be someone from replit team can tell something helpful.