CivArchive
    Wan SCAIL-2 Segmentation Control - v2.2
    NSFW

    Wan SCAIL-2 Segmentation Control

    Features:

    • Image Analyzer

    • LoRA Support

    • Interpolate | Upscale | Color Match

    • Color Correction

    • Image Sharpener

    • Sage Attention

    • Choose between 2 Samplers

    • Background Remover (RMBG) to keep the background of the input video

    • SCAIL-2 Identity Tracker

    • Load an alternative audio file for the final video output

    • Installation Paths & Download Links

    • Well-organized

    • Interpolate the Input Video

    SCAIL-2 Identity Tracker:

    For Multi-Character set "object indices" to "nothing" (empty field, no value). Then set the SCAIL-2 Identity Tracker to "Point" and select your characters by clicking on them in the image below. You can also try "Box," but if there are more people present, this could lead to problems. Use a start image similar to the one in the video.

    Note:

    If you have any questions, please first read the information in the red boxes within the workflow.

    Additional options are available within the subgraph. Click the icon in the top-right corner of the Main Settings node to open it. Help is available by hovering your mouse cursor over the values inside the subgraph.

    The workflow offers two samplers. Both deliver similar results. Since I like both, and for testing purposes, the workflow allows you to easily switch between them. Personally, I use SCAIL Auto Extend, This one seems to have no or fewer color shifts. And doesn't need the "Color Match" option.

    Wan SCAIL-2 is not perfect, but it delivers good results in most cases. If you encounter issues, setting a new seed or switching the sampler usually helps.

    Description

    FAQ

    Comments (19)

    oo7mJun 14, 2026· 3 reactions
    CivitAI

    There is an error in loading the workflow in latest Comfyui, please check.

    Thalion7
    Author
    Jun 14, 2026

    I have. What kind of error do you have?

    oo7mJun 14, 2026· 1 reaction

    Loading aborted due to error reloading workflow data

    TypeError: (v || []).reduce is not a function

    This may be due to the following script:

    /extensions/comfyui-lora-manager/loras_widget.js

    oo7mJun 14, 2026

    Quickly used Claude to fix it https://pastebin.com/WExgMAA3 Yet to test it

    oo7mJun 14, 2026

    Can confirm that the workflow at https://pastebin.com/WExgMAA3 works fine

    oo7mJun 14, 2026

    I also had to upgrade comfyui to the latest nightly for it to have all the nodes working

    oo7mJun 15, 2026

    I also see that with this workflow, there is a drift in the facial consistency as compared to the original workflow.

    Thalion7
    Author
    Jun 15, 2026

    @oo7m That's interesting. Thanks for the info. I'll have to compare them.

    Thalion7
    Author
    Jun 22, 2026

    @oo7m I believe the problem with the facial consistency is related to the starting image. The more the beginning of the animation deviates from the input image, the greater the problem. Don't you think?

    mayayaheeJun 14, 2026· 2 reactions
    CivitAI

    The real SCAIL strength here is multi character, could you please provide some nsfw examples?

    kudon44Jul 7, 2026

    I'd be curious too, because it seems to perform very poorly when I attempt more than one character and I've noticed a distinct lack of anyone posting multi-character outputs that aren't just them basically standing there. It's proven particularly rough, or maybe even impossible, for NSFW materials to swap two or more characters in one generation.

    kudon44Jun 25, 2026
    CivitAI

    How do loras work with this? Is it just Wan 2.1 lora or do 2.2 also work? And if so, only low lora, no high?

    Thalion7
    Author
    Jun 26, 2026

    I updated the workflow. It's using Wan 2.1, but maybe the Wan 2.2 Low LoRA will work as well.

    bvgaJul 7, 2026· 1 reaction

    I've had good results with certain Wan 2.2 high noise LoRAs.

    Thalion7
    Author
    Jul 7, 2026

    @bvga would you like to post them to my workflow?

    pencilcheckJun 28, 2026
    CivitAI

    can you create a runpod template for this? thanks also I don't know how to use this, where is the node to upload the image and also upload the video?

    Thalion7
    Author
    Jun 28, 2026

    Sorry i can't. Did you download the workflow for ComfyUI? The Nodes are on the left side.

    pencilcheckJun 30, 2026· 1 reaction

    @Thalion7 I did downloaded but I don't see it that's why I asked

    akacicala385Jul 26, 2026

    @pencilcheck Use a rope and hang fgt

    Workflows
    Wan Video 14B i2v 480p

    Details

    Downloads
    1,397
    Platform
    CivitAI
    Platform Status
    Available
    Created
    6/14/2026
    Updated
    8/25/2026
    Deleted
    -

    Files

    wanSCAIL2Segmentation_v10.zip

    Mirrors

    wanSCAIL2Segmentation_v10.zip

    Mirrors

    wanSCAIL2Segmentation_v22.zip

    Mirrors

    wanSCAIL2Segmentation_v20.zip

    Mirrors

    wanSCAIL2Segmentation_v22.zip

    Mirrors

    wanSCAIL2Segmentation_v20.zip

    Mirrors