I made a simple prompt scraper for my images.

I use a lot of random pools for my slop generation. I wanted a way to be able to try to recreate or modify an image based off the prompt. I used chat GPT to write some python code to get it done, and does it in batches, if there are a few prompts you want to scape.
i was also tired of the rough way meta sites showed the meta.



pretty sure you will need you input your own paths for the directories, but right now it will pull the images from my "images to rip" folder and output a text file with the positive and negative meta data only into "ripped" folder along with the image that it came from to not get duplicates. looks like it will overwrite images with the same name when moving, so be aware of that.
you can just run the .py file and it will tell you that its done.
still no idea what I'm doing but its neat that I am getting something functional.

updated to work with new comfyUI stuff.