Struggling to build "chat" logic for OpenAI API in Telegram

Hey there!

I’m trying to do the following, but I’m failing:

  1. I want to create a bot for Telegram with the ChatGPT API, which, when addressed, sends back a set of emojis that “hint” at the name of a movie.
  2. In response, the user writes their guess and receives an answer whether it’s correct/incorrect.

I don’t understand how to implement the logic of such a “chat”.

What I’m doing now:

  • I’m using the Telegram module Watch Updates.
  • The user triggers it and launches the generator on the OpenAI side.
  • I send the received result of the ChatGPT module to the “send message via Telegram” module.

But then, I don’t understand how to make it so that the subsequent user message is considered as a response, rather than triggering a new question.

In general, I don’t understand how to implement the concept of “ChatGPT” (in terms of “chatting”) in a third-party messenger through Make.com.

I tried making a router with the following logic:

  • The question is asked as a regular message
  • The user answers it via reply.
  • Then in the attributes of the result of the module, the original message appears, which would partially solve the problem.
  • I would filter messages with and without a reply through a router and the logic would work.
  • But the problem is that I can’t select this variable “reply” from the list of variables, it’s not there (although again, it’s clear that these data are transmitted).

Looking for your experience, thanks!