CivArchive
    Сonsistency characters | Generate characters only by image and prompt, without character's Lora! | IL\NoobAI Edit - v0.4
    NSFW
    Preview 108368044
    Preview 108355653
    Preview 108355763
    Preview 108356157
    Preview 108356165

    This is a test workflow that allows you to use the IL\Noob model as Flux.Kontext\Qwen_Edit to generate a character image from a Reference. It works best with the same model as Reference. You also need to add a character prompt.

    Attention! The result depends greatly on the seed, so experiment.

    I really need feedback and advice on how to improve this! So if anyone is interested, please share your thoughts on this.

    Description

    * Removed "everything everywhere" and made the relationships between nodes more visible.

    * Support for "ControlNet Openpose and Depth"

    * Bug fixes

    FAQ

    Comments (9)

    linelinelNov 13, 2025
    CivitAI

    That's really cool, thanks for puttin everything together!
    Do you think, using this logic, it'd be possible to do the same for a given environment/scene?
    Like, have a base image of character in a given environment, and be able to generate other framing/posing in a similar looking scenery

    Slavikson
    Author
    Nov 13, 2025

    Yes, someone on reddit wrote that they modified it a bit to do what you described. In the next version I will add native support for working with backgrounds/scens.

    linelinelNov 13, 2025

    @Slavikson ah nice, thanks for the quick answer!
    Also, in your v0.4, the depth image appears to be cropped (i'm giving 9:16 ratio, and the output depth is 1:1) so i end up loosing the bottom and top part of my ref image
    I can change the resize mathod from fill/crop to pad or keep proportion, but then the generated images are glitched

    Any workazround?

    Slavikson
    Author
    Nov 13, 2025

    @linelinel for "Apply ControlNet" to be applied correctly the image must be adjusted to the canvas size. You can try from the other side - change the working canvas. This is the bottom right "🔧 Image Resize" in the "Ref" group and "🔧 Empty Latent Size Picker"

    linelinelNov 13, 2025

    @Slavikson Ahyea I see
    Thanks a lot mate!

    TKuro1988Nov 17, 2025· 1 reaction
    CivitAI

    I like the idea but it's way too complicated to read and understand. Also some models aren't linked and hard to find or obsolete or maybe you renamed them.

    Slavikson
    Author
    Nov 17, 2025· 1 reaction

    I agree that it is complicated, and the situation will not improve in the next version. I left the link to the necessary models in "Link", if there is no link there, you can use any similar model.

    AnshoJan 31, 2026
    CivitAI

    i'm having problems with constant errors, does the workflow need a update because of the recent update?

    Slavikson
    Author
    Jan 31, 2026

    Maybe. Can you tell me which nodes stopped working?

    Workflows
    Illustrious

    Details

    Downloads
    1,575
    Platform
    CivitAI
    Platform Status
    Available
    Created
    10/31/2025
    Updated
    5/4/2026
    Deleted
    -

    Files

    OnsistencyCharactersGenerate_v04.zip

    Mirrors