Add native ComfyUI API support. Generate images with a local ComfyUI server

ComfyUI is a free open source method to use AI to generate images on your PC.

The only way to use it in SCM is via webhooks, however we can promote ComfyUI to proper native support.
ie You select via a dropdown, with model detection etc

So now we have proper support for it.

How to connect ComfyUI

Go settings > API logins > AI image

Most values are pre-filled (including URL)

You can click on the dropdown to get a list of all checkpoints installed. No copy paste required.

Thats it.

How to use ComfyUI

You can use it in 2 locations.

Image downloader

The image downloader got reworked. You just select the ‘AI image’ chip.

Select ComfyUI from the dropdown

Then type in a search and press enter.

Results appear below.

Full select and export controls as well.

Page Pipeline

Inside the Inserts tab.

Under Image

Enable AI image and select ComfyUI.

Now the Page Pipeline will create images usin ComfyUI.

Done.

Simple connection and go!

1 Like

Thank you so much. This is a big milestone to be able to generate the pictures locally. Another big one is missing and should be easy to implement if you can in the next update. Local ollama support. I don’t like to have both running LMStudio and Ollama for other uses.

1 Like

I can look into it!