Earlier, generating images on text apps wasn't possible, you'd have to select an image generation model like Dall.E 3 which would just generate images. Now, you can use any text based LLM and enable image generation for it. 

The only caveat is that an extra prompt will be used for every request to check whether an image should be generated or not. This can slow down responses. I'm using Dall.E 3 right now as the model for image generation but may switch to Dall.E 2 or something else in the future. Pro and Pro Plus users will have the option to select which model they want to use for this image generation feature in the future.

To enable this option, select "Enable Image Generation" from the advanced settings from your app. I'm also working on adding image to image generation functions (also called impainting) which will let users upload an image, pass a prompt, and get a modified image.