Anima workflow
GUIDE HERE a bit outdated, but should give general information on how the workflow works.
First version of the inpainting workflow came almost 3 years ago for 1.5. It´s been a long ride 🥳
13/07/2026 - v1 Beta anima
Added ability to do full image inpainting. This has been requested multiple times.
Added ability to use edit loras with cosmos reference. See Anima edit and Cosmos Reference . The node and lora are necessary for using the edit. If you see shading/shadow of original image, raise "Expand mask area".
The workflow is a bit more messier than 0.9, as bypasser/muter does not work with subgraph. Did not bother to figure out how to clean it yet, hence the beta.
With edit you can do simple edits to the images, limited by the lora. Like changing clothes, adding stuff etc... Other than what you want to edit, generally stays the same.
Edit enabledDenoise: 1prompt: add cat ears, animal ear fluff,cat ears,blue sky,masterpiece, best quality,Let me know if you find something funky.
26/05/2026 - v0.9 initial release for anima
Converted XL workflow to work with anima. It works pretty much same as XL one, guide is relevant on that part.
Inpaint workflow XL V2
GUIDE FOR V2 WORKFLOW IS LOCATED HERE
06/12/2025
-Added ability to control mask opacity.
-Changed sampler to prevent problems with generation.
-Adjusted guide.
04/12/2025
-Added ability to do colormatch. Changing colormatch setting to 1 will try to match the color of the image.This is good if using VAE that changes color between encode/decode cycle or using different lora/model.
Older:
-Added differential diffusion
-Cleaned the workflow
-Created guide which can be found HERE
-This should work with 1.5 too, did not test.
Inpaint workflow XL V1.1
Due to request updated to work with XL. It is somewhat barebones compared to 1.5. Does not have segmentation or controlnet models currently.
I have a bit outdated comfyui, let me know if it is throwing some errors.
Description
-Added colormatch.
Using 1 on color match will try to match the color to the original image.
-Added note for all the settings
FAQ
Comments (18)
In my recent runs, the inpaint workflow renders the masked area as a solid green overlay instead of properly blending the inpainted result.
You have helped me a lot in my eternal search for the ideal. Thank you, Great man
I don't know why I didn't leave a comment earlier.
(～￣▽￣)～ glad you like it
It would be great to be able to resize the latency space. Working with the square limits our possibilities. I have not yet figured out how to bring the inpaint from Forge closer to Comfy.
When i started building the inpaint workflow over 2 years ago, my idea was to replicate how inpainting worked in A1111. It has come long way since then and i personally still use slightly edited version of the workflow every time i inpaint. I don´t exactly know how it differs from current forge as i have not used it in year or so.
Main reason why i created it to cut square was to have better results for the inpainting. For XL for example you are going to get good results mostly on specific resolutions (1024x1024) as example. If it uses the cut resolution on generation, it could cause distortion and consistency issues with higher denoise and the results will not be as good as the generation is done on the lower resolution image (unless the cut area size is larger than the 1024x1024 or which ever resolution you are generating with).
If you want to inpaint large area on image, at that point inpainting with the full image might be better choise. Getting detail on images was always the focus for this workflow.
Thank you for the comments (～￣▽￣)～
@PotatCat 💜
I don't know if it's just me, but ComfyUI can never find the "Images to RGB" node not matter how many times I reinstall the Impact Pack
You could try remaking the node
Also isn´t that node from WAS node suite
e: yeah it is from this pack: https://github.com/ltdrdata/was-node-suite-comfyui should be able to find with comfyui manager
Has something recently changed in how the workflow interacts with ComfyUI? It doesn't seem to be inpainting in the mask any more but modifies a broader chunk of the image in a negative manner. I thought maybe it was just my edits so I redownloaded the basic workflow and still having the same problem.
See if you disable nodes 2.0 if it works
This workflow is really great and gives me better results than Krita Diffusion, especially for color matching. I was wondering if you could maybe provide an Anima version as well, if possible. I tried adjusting it myself for Anima, but the images (masked regions) kept coming out blurry.
Thanks for the comment. I have made one for anima, it is currently messy and so on. I try to find time to make it usable by others (～￣▽￣)～
@PotatCat Thanks mate, that would be great!
@IceB2 Edited the XL workflow to use anima, was pretty painless (～￣▽￣)～ Did not test it that much, let me know if you see any problems.
@PotatCat Much appreciated, I'll get to it when I have time.