1. Uses ComfyUI-nunchaku for load optimizations

2. Uses comfyui-image-prompt-loader for loading either local image or civitai image with or without metadata, copies prompt/seed/step values into workflow

3. Runs image through Depth

4. HighRes Fix is then run

5. Outpainted 3 times incrementally instead of a large singular outpaint

6. Ultimate Upscaled

7. 2 x Detailer nodes (Face/Person/whatever extra detailing you want)

Use Case

Grabbing an image you scraped with either civitai, gallery-dl or other tools for AI gen images that contain prompts, so you can:

• Reproduce the image

• Tweak or iterate in a slightly different way from original image

• Just grab prompt only from original image

• Just use original image for depth

• Expand and upscale original image

• All of the above in one single flow or individually if wanted