Photoshop to Stable Diffusion (Single Node, updated)

A very quick update to my previous Photoshop to ComfyUI tutorials, which, since Nima's updated their nodes, needed a bit of a refresh.

Everything else can be installed via ComfyUI Manager by dragging and dropping the workflow file!

(it's the diffusion_pytorch_model safetensor file, rename it to something more memorable)

Timestamps:
00:00 - Intro
00:42 - Workflow Overview
01:48 - Photoshop to ComfyUI node install
03:33 - Set up the Connection
04:02 - Starting the Generations
05:47 - Enabling the Optional Relight Group
06:53 - Quick Comping Use Case
07:38 - Outro

#stablediffusion #photoshop #comfyui #stablediffusiontutorial #photoshoptutorial #comfyuitutorial

[jfosterdesigns]

Fantastic! Thanks so much for incorporating Nima's node. Hopefully, he'll incorporate your workflow and nodes into his plugin!

[LinhLe-ibgi]

i don''t connect PS with comfyUI . Image in Photoshop not show in ComfyUI, how to use it ?

[tepa]

Hey Andrea, i like your content, specially the professional creatives side that sets some standards so people can have control for a change, rather than having AI spit out whatever it wants. However following this tutorial i am having issues with popup window not showing up. - I have also followed that suggestion you left in the other video(description) where you need to install requirements in the "comfyui-popup_preview" folder, but still its not giving a pop up window. - I am on windows 10 btw.

[saberkz]

Looks cool thank you for this video

I have a request can you make similar but by drowing like anything in the picture then it get apply into SD preview ? For example a product photo then we drowing a sun and sky in canva and adding props etc then SD apply it on the frame

[jccluaviz]

Seems popup its not working for me. It's not showing neither in Photoshop nor in any other window. What i'm doing wrong ?

Many thanks.

[albertrc]

Hi! First of all thank you so much for your content. I've already used the workflow in Windows and it runs with no errors but now In a Macos it appears the error:

When loading the graph, the following node types were not found:
PhotoshopToComfyUI

I've tried to install it via git and manually but it's still not working, any idea of why it's not installing correctly?

Thank you in advance!

[SejalDatta-lu]

Recent subscriber here - great content!
I keep getting this error. Do you have any ideas?
"Error occurred when executing ImageScaleToMegapixels:
NoneType' object has no attribute 'shape'"

I've ensured that I've got all of the same models as you, although my picture in Photoshop is a different size.

Thanks

[HanjLaoye]

The following node is missing: PhotoshopToComfyUI, how to solve it, thank you!

[adberf]

Keep going Andrea, keep creating content. You are the best. 🤍

[coreypan]

helpful content, thx!!! But theres no a popup show up in photoshop, could anyone tell this ?😭

[CGFUN]

Hey man, since u intersted in SD, have you took a look to tydifusion inside max ? its amazing and easy to use then nodes.

[SejalDatta-lu]

How do you keep the image produced by ComfyUI as similar to the images in Photoshop as possible.

e.g.
In the case of the female mode in your video, or even a landscape background, comfyui makes changes.

While you might be able to solve the face using a face swapper, it might be a deal breaker for products or clothes that need to be exact.

Any ideas?

[agusdor]

Will this plugin works if I use comfy trough a cloud service like runcomy or runpod? Ty!!!!

[dimitrishow_D]

I've been trying to make ai create my cartoon character in diff poses or with diff clothes...I must say everytime I try .. eventually I end up just doing it myself...

It might be good for creating stuff out of nothing realistic persons or anime but it can't seem to work in my style since there is no real reference to it..

I tried training my own model but that also always ends up looking like a mentally handicapped person is trying to make my stuff..

So yeah it's fun but it's not good enough to create real specific stuff quickly

[DarioToledo]

Not sure that my 4gb of VRAM are gonna handle that 😅