From the examples of Wolfram, seems like they exported the visual as an image file and asked ChatGPT to render the image directly.
I am curious if it is possible for my plugin to return some raw html/js/css and ask ChatGPT to render it in the chat session?
It seems unlikely at this moment. But is OpenAI considering opening up this possibility? Thanks a lot!
The chatgpt interface doesn’t render html or css yet, maybe one day.
But for now, you can render Markdown. You’ll just have to play around with your setup to get it to consistently respond to the user with Markdown.
Here is an example, I prompted GPT specifically so it would output the markdown code without code delimiting ticks:
Did not find solution, so I’ve built my own (of course) 
codegpt-16db3.web.app
It will render simple HTML/CSS/JS, has versioning and sharable links.
