Sign In
My ComfyUI workflow

Updated Workflow

This is a short update on my previous workflow. Although the basic image generation pipeline is more or less the same, I decided to make some changes anyway.

So basically the pipeline looks like this:

  1. Generate a sketch (1st Pass group)

  2. Fix faces (FaceDetailer group)

  3. Fix hands (Hand Refiner group)

  4. Upscale (2nd Pass group)

  5. Save image and embed prompt metadata

Feel free to use this workflow. I really hope you find it helpful <3

Custom Nodes

ComfyUI-ImpactPack

GitHub Repository

Quite a decent node package for various needs.

In my workflows this node pack is used mostly for post-processing and detailing. I think this is a must if you really need to improve the overall image generation quality.

To be more specific here are Face Detailer and Hand Refiner flows:

Also I really want to mention "ToBasicPipe" and "FromBasicPipe" nodes. These beauties just make my life easier. Just try them ;)

rgthree's ComfyUI Nodes

GitHub Repository

Currently I use this node pack because it has a life-saver node "Context Big". Similar to ImpactPack's BasePipe nodes, it allows to comb all these eldritch noodles between ComfyUI nodes and reduce visual complexity of a workflow.

SD Prompt Reader

GitHub Repository

This is the pack I decided to use firstly because of the ability to embed CivitAI-compatible prompt metadata in your output image. Then I found out that it has way more features XD. But now I use SD Parameter Generator, SD Lora Loader and SD Prompt Saver nodes.

Save Image with Generation Metadata

GitHub Repository

This one has some hidden features I use on regular basis: "Seed Generator" and "String Literal". These two nodes add some good UX detail for me :)

Cute Comfy

GitHub Repository

This cutie make the overall UX more... cute?:) Really, this one is a personal favourite <3

5

Comments