Try MLLM Guided Image Editing (MGIE)

MLLM-guided Image Editing (MGIE) - a Hugging Face Space by tsujuifu

Here you can try out the model being used by Apple in their new Siri image generation tool in iOS 18. This is a kinda clever approach where they take a simple user prompt and customizes it to be more intricate and oriented toward the the input image. This leads to more consistent and predictable edits of existing images.

One of the hardest parts about AI editing with images is figuring out what kind of prompt to use to get the desired effect. This seems to bridge that gap by allowing you to add very very simple instructions and then get back in an intricate description of what needs to happen on an image customized to that image