Generating content via assistant

Hello everyone,

I’ve been working on generating trivia questions for my mobile app. Here’s what I’ve done so far:

  1. Created an assistant.
  2. Crafted an ‘instructions.jsonl’ file to define the desired output format.
  3. Fed the assistant with a ‘questions.json’ file, which I’m currently using in my app, to establish a pattern.
  4. Developed a Node.js app to retrieve the output.

I have a few questions:

  1. Is it advisable to create an assistant for such a purpose?
  • If yes,
  1. Is a function necessary? I’m currently using it to clearly define parameters like the ‘amount’ of questions to be generated, the ‘category’ of the question, and the ‘language’ of the questions and answers.

I’m currently facing an issue where the ‘submit_tool_outputs’ message interrupts the conversation flow, but I require the output directly in JSON format (as specified in the instructions). Can anyone provide guidance on this matter?

Thank you!