How do you make an assistant generate pictures after collecting information about the user?

So I am creating a new assistant that is to collect information about the user. Once that information is collected I want it to create 3 different images based on that info.

I am working with the playground right now.

The bot is collecting the info and even saying that it is going to generate the images, but it never does that. Is there something specific I need to do in the prompt to tell it to go to DALL-E mode?

You would need to write a tool specification and function yourself that calls upon the images API with the appropriate set of parameters.

You already specified the model to use when you made the assistant.

1 Like

Awesome. Ok, I am new at this. Where can I learn more about tool specifications and functions?

Found the cookbook… looking in there…

API reference and documentation, found in the forum sidebar, each have their section about assistants and tool use.

You can also start with the foundational info linked under chat completion for functions.