merserk/comfyui-ultimate-face-fix — explained in plain English
Analysis updated 2026-05-18
Automatically detect and repair one or many faces in an AI-generated image.
Fix faces in anime, illustrated, or realistic style images using a matching generation model.
Build a custom face-crop enhancement pipeline with the Extract and Process nodes.
Use ready-made example workflows to fix faces with SDXL or Krea 2 Turbo.
| merserk/comfyui-ultimate-face-fix | 0c33/agentic-ai | adennng/stock_strategy_lab | |
|---|---|---|---|
| Stars | 14 | 14 | 14 |
| Language | Python | Python | Python |
| Setup difficulty | moderate | hard | hard |
| Complexity | 3/5 | 4/5 | 4/5 |
| Audience | vibe coder | developer | researcher |
Figures from each repo's GitHub metadata at analysis time.
Requires ComfyUI with a compatible image generation model already connected.
ComfyUI Ultimate Face Fix is a set of add-on tools for ComfyUI, a visual, node-based program used to generate images with AI. This add-on focuses specifically on repairing faces in generated or existing images, whether there is one face or several in the same picture. It works by first detecting each face, then repairing a cropped square around it using whatever image generation model you already have connected, and finally blending the repaired face back into the original picture using a precise mask so the edges do not look obvious. It supports many popular image generation models, including different versions of Stable Diffusion, FLUX, and several others, and it can handle anime, illustrated, and realistic art styles. Installing it is done through ComfyUI Manager by searching for its name and clicking install, which also downloads the extra face detection models it needs automatically. There is also a manual installation option for people who prefer to clone the project directly and run a setup script themselves. Once installed, the tool adds several new nodes you can drag into your workflow. The main node offers many adjustable settings, such as how many faces to fix, how strongly to change each face, what size to resize each face crop to before processing, and how to blend the repaired face back in. There are also optional connections for more advanced face detection and masking tools if you want finer control. For people who want more customization, the project includes separate extract and process nodes so you can add your own image editing steps to the face crop before it gets repaired. There is also an automatic denoise helper that looks at detected faces and suggests a reasonable repair strength based on their size. The project ships several ready-made example workflows to help new users get started quickly. The code is released under the AGPL-3.0-or-later license, while any downloaded model files keep their own separate licensing terms.
ComfyUI nodes that automatically detect and repair faces in AI-generated or existing images using your own connected generation model.
Mainly Python. The stack also includes Python, ComfyUI, YOLO.
You can use and modify it, but if you run a modified version as a network service, you must share your source code under the same license.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly vibe coder.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.