Images from sites like Tensor.art, NovelAI, CivitAI, or the desktop application ComfyUI, all have EXIF data that is readable through the application, but it would be nice to be able to query through this data and or parse the prompts as descriptions automatically.