Comfy-UI Most ADV Setup

Download it, Extract it in the Folder that you want, run it like Comfy-UI, then drag and drop, Nico Setup v???.json on Comfy-UI.

Now let’s get started with explication

Inside you will find 6 Block (Group), Helper, Control Room, TXT 2 IMG, IMG 2 IMG, Upscaler, Paint.

I assume that you brain know what is TXT 2 IMG, IMG 2 IMG, Upscaler and Paint.

So i will explain the uncommun stuff, (Helper, Control Room).

Control Room

The Contol Room, control everything, if you didin’t get it.

• Block Activation : It’s on the right and self explanatory.

• Load Image : It’s quit special because you need to drag and drop the on it, to load it. You can also creat mask and edit the image (Great for heavy Painting). I wont explaine this node further, find out yourself.

• Loader : Self explanatory.

• Lora Stacker : Self explanatory.

• Helper Weight : It go from 0 to 100, but you will only use it between 5 and 40 because of distorsion. This node will be explained when i talk about the Helper Block.

• Helper Select : Same it will be explained when i talk about the Helper Block.

• [ N ] Prompt : Negative Prompt, if someone ask me, (Nico what do N prompt mean ?) i will jump from my window. I have embeded the Universal Prompt +++ (From Model-EX) inside the setup, so for the humanity sake don’t add N-Enhancer Tag like (ex: Bad, Very Bad, Bad Details, Bad Quality, Bad Fingers, etc…) Please…

• [ P ] Prompt : Reapeat what i say for [ N ] Prompt with the contrair of Negative. Universal Prompt +++ Is also embeded here. Don’t add Enhancer Tag like (ex: Details, Quality, 8K, Good Fingers, etc…) Please… For God Sake.

• Color Correction (Color Presision) : This is all the color nodes at the side of [ P ] Prompt. Basicaly if you put multiple colors in your [ P ] Prompt (ex: White Skirt, Blue Shirt, Red Hat, Purple sky etc…) You will not have the colors at the right place or not even have it. Color Correction it’s solving the probleme, if you absolutly need a (Red Shirt), you will go to the red node and add (Red Shirt), you need to add (Red) at each prompt tag.

Helper

Is a dictionary of all Instant Lora. Beside every Instant Lora you will find it’s name and a number. Put the number inside the Helper Select, to use this Instant Lora, and the Helper Weight it’s the Lora Weight, it’s set at 20 by default because it give the most flexibility without lossing the Lora.

You also have an image of the Instant Lora with a Text on top, The text will be added to the [ P ] Prompt when the processing is runing, so there is no need to copy and peaste it in [ P ] Prompt.

4 Blocks

• TXT 2 IMG : Have a double sampler to get the best of the Model on the first try. Remember to add a path to the Auto-Saving image node because the block will not work without it.

• IMG 2 IMG : Will start with a image resize of 1024 pixel on the longer side (That dosen’t mean tones of details, except if you image is 4k), then it go though 3 heavy sampler for correction, then inside of the Tiny Upscaler that will add 200 pixel to the longer side, to give more room for understanding for the 2 next soft sampler, those sampler will correct mainly light and small details. Remember to add a path to the Auto-Saving image node because the block will not work without it.

• UPSCALER : Will start a big upscaler, set by default at 2048 pixel on the longer side, (You can chose the size), then it go inside of a Tiled Sampler that will add the details to you enlarged image (info: The higher the image size, the Tiled Sampler will need a higher denoise, not recommended to go above 0.55). Remember to add a path to the Auto-Saving image node because the block will not work without it.

• PAINT : In paint you will only have the option to Outpaint, in the outpainting node and see the mask you created because the mask creation is actually done in the image loader when you click Edit, and just like a said before, figure it out yourself on how to use the Editor tab. And again don’t forget to add a path to the Auto-Saving image node because the block will not work without it.

Youtube

If you are a content creator, and you want to show case this setup, you don’t need to ask for permission. I upload it for the community, everyone can do what ever they want with it.

your free to make a video and it will be a pleasure for me 🙂

Participate in Helper

If you want me to add a character for the next update here are the contidition :

• 3 Image

• White Background

• Inside the image you need to put the full body with clear vision, hands, face, hair, body proportion and default dress. ( If you can’t squeeze in the lower leg it’s not a problem).

• 3 Similar Image, in different pose ( Don’t make a pose with with the back turn).

And i will queeze your name beside the character name, (For flexing).

Update and Motivation

Im Working hard on this shit (I use it so much) so update will come often (Very Often).

The main source of my motivation to continue publish this to you guys, is to see you using it (It make me very happpy), so send me a screenshot of you using it, it will put a smile on my face, and commentary are no excaption 🙂