Automated DALL·E 3 and GPT-4 Vision Recursive Image Generator ("Make it more" trend)

[acoustic radish]

Leverage the DALL-E 3 and GPT-4 VIsion APIs to generate and image, have GPT-4 Vision describe it, and use the description to generate the next image.
It's like Gartic Phone/Telestrations for AI Images!
Enjoy creating amazing results!
Link: https://image-recursor.web.app/

[snow swift]

cool idea, asking for an API key is... not great practice.

[acoustic radish]

Yea I agree it's not the best solution - the website is mainly a side project and I didn't want people to pay me or have me incur costs for no reason and couldn't think of any other way.

[dense lagoon]

so it makes better results over time?

also, considering this is some hobbyist project for you consider open sourcing it, i don't mean to be rude but I don't want to give a random person my api key

[acoustic radish]

Oh yea will do this!

[acoustic radish]

Not better, just different. For example:

(It’s a gif)

[acoustic radish]

https://github.com/PallavAg/image-recursor

Automated DALL·E 3 and GPT-4 Vision Recursive Image Generator ("Make it more" trend)

New UI!