CivArchive
    Flux Redux Workflow - Kinda Sorta IP Adapter :) - v1.5
    Preview 41421632
    Preview 41421846
    Preview 41421601
    Preview 41421600
    Preview 41421630
    Preview 41421842
    Preview 41421844
    Preview 41421845
    Preview 41421843
    Preview 41421629

    Two Flux Redux workflows!

    One combines your input prompt (for composition) with an input image (for style) and creates an often-unexpected mutant hybrid!

    The other merges style/composition from any two images with no text prompt input.

    Enjoy!

    Description

    Now with 2 workflows!

    Simplified the original text + image Redux workflow, and added a separate 2 image mixer workflow.

    FAQ

    Comments (6)

    _kaidu_Nov 22, 2024· 7 reactions
    CivitAI

    Hi,

    I don't want to spam your workflow with advertisement, but just in case you are interested: I made a ComfyUI plugin that allows for changing the strength of Reflux and balancing out Reflux and user prompts. It works really well on all my experiments.

    Here is the github repository to the plugin:

    https://github.com/kaibioinfo/ComfyUI_AdvancedRefluxControl/

    EnragedAntelope
    Author
    Nov 24, 2024

    Hi, you've come up with a good method, this is unique and often does produce different/better results than most of the other conditioning tricks we're seeing.

    I would suggest:
    1) Updating name of repo to include "Redux" so people can search for Flux Redux and find it (plus "reflux" is quite uncomfortable medically :) )

    2) Update your node names to not conflict with Kijai/other repositories - make them unique please

    3) Include on Comfy Manager

    These things will definitely reach more people and also avoid node name conflicts within the Comfy ecosystem. Thanks!

    k8kissDec 4, 2024· 1 reaction
    CivitAI

    Great contributions!! Do you have a Flux default text to image and image to image workflows that you always start with for photorealism?

    EnragedAntelope
    Author
    Dec 4, 2024· 1 reaction

    Thanks! Check my models page https://civitai.com/user/EnragedAntelope/models - I usually use the Clip-L/T5 workflow for Flux for text to image.
    I usually don't do straight image to image, I have some inpainting workflows you can check out for those. Enjoy!

    k8kissDec 5, 2024· 1 reaction

    @EnragedAntelope Awesome!! I will try that workflow that you suggested. I tried one of your image embedded workflows but it has lots of floating empty windows and it was messy and hard to follow. Maybe there was an issue with the embedding. Will let you know how it goes!

    EnragedAntelope
    Author
    Dec 5, 2024

    @k8kiss very possible that since i posted, nodes have changed due to comfy updates. Some trial and error may be needed!

    Workflows
    Flux.1 D

    Details

    Downloads
    531
    Platform
    CivitAI
    Platform Status
    Available
    Created
    11/22/2024
    Updated
    5/13/2026
    Deleted
    -

    Files

    fluxReduxWorkflowKinda_v15.zip

    Mirrors