black-forest-labs

/

FLUX.1-dev

like 4.5k

Yes, it can definitely be used for img2img when using ComfyUI. Just google "Flux img2img Workflow ".
As for using the API, there is a demo on https://fal.ai/models/fal-ai/flux/dev/image-to-image/playground

The JSON should look like this:

{
  "image_url": "https://fal.media/files/koala/Chls9L2ZnvuipUTEwlnJC.png",
  "prompt": "a cat dressed as a wizard with a background of a mystic forest.",
  "strength": 0.95,
  "image_size": "landscape_4_3",
  "num_inference_steps": 40,
  "guidance_scale": 3.5,
  "num_images": 1,
  "enable_safety_checker": true
}

Couldn't find a demo on Replicate but you should be able to adapt it from the JSON I think.

How to use img2img with diffusers library?

Sure, but I'd rather have img2img's pipline