Sign In

Metadata editor, ComfyUI outputs into Civitai with resources. (Abandoned)

9
Metadata editor, ComfyUI outputs into  Civitai with resources. (Abandoned)

I made this simple tool to inject or replace meta data into an image, is made mainly for add loras and checkpoints so Civitai recognize the resources that you used, is still kind of basic but if someone likes it I can make new versions with more functions, so far works for windows, any feedback is welcome, I'm not that good at coding so may not be perfect. I guess this can be used to lie in what you used but pls don't do that is for exactly the opposite give credit to someone else work (loras or something).

Download Here.

Git repo: https://github.com/JAdrianPZ/MetadataEditor/tree/main?tab=readme-ov-file

So far works for checkpoints and loras.

You can also use it for Flux is still not perfect for that tho, seems to no catch the checkpoints, but is ok for loras.

Is a tricky workaround here:

FLUX: The flux one is a little tricky, works fine with only loras but not checkpoints so your images may not show under that category, others models work fine but im gonna work for this one, so far if you want to add lora and checkpoint you can do it manually with de edit data yourself (the box under the image) and follow this example with your info:

A detailed 4K anime-style image of a  man holding a sign that reads: 'I made this with ComfyUI but Civitai know what resorces i used', he looks happy, the back ground is a cyberpunk city 

Negative prompt: 

Steps: 30, Sampler: Undefined, CFG scale: 3.5, Seed: 1944373770, Size: 1024x1024, Clip skip: 2, Created Date: 2024-08-31T22:37:48.7691153Z, Civitai resources: [{"type":"checkpoint","modelVersionId":691639,"modelName":"FLUX","modelVersionName":"Dev"},{"type":"lora","weight":1,"modelVersionId":782014,"modelName":"David Martinez","modelVersionName":"v2.0 Flux"}], Civitai metadata: {}

Like this:

I know that out there are more complete options but this one is easy to use and straight to the point.

Usage: Just select or drag and drop an image, then add the info for prompts and select the model you used and the loras, if the image contains metadata is going to show it in the left side(don't work for some images), but still is gonna inject the new one. May still not have all samplers but tell me if you find this useful so i continue working on this.

9

Comments