CivArchive
    K3NK 4llinOne - NSFW Helper - Wan 2.2 Video LoRa - K3NK - Low noise - v1.0
    NSFW

    ☕KOFI https://docs.google.com/forms/d/e/1FAIpQLScltzaQXTF2VFWRWVaqPSzJXYyNBF9hUQOLI8aBCyLrpCM1_w/viewform?usp=dialog

    💢 I appreciate feedback but, really guys you shouldn't come here to tell others to not support me because you get bad results, I'm not Amazon, I'm not "selling" the lora, early access is a way to support our work so we continue doing it, some of you are crazy really... You can check the metadata/workflow of the videos I uploaded.. Please upload your videos with the workflow then complain, thanks ! 💢

    Ultimate NSFW Helper LoRa for Wan2.2 I2V

    • Trained on diffusion pipe with a dataset of 5131clips of various resolution, 1sec to 3sec or 5sec (3hr to complete 1 epoc, num repeat = 1)

    • Optimizer/learning rate: automagic

    • both model trained at 211 bucket

    The lora contains multiple of the datasets of my existing loras and some additional ones, i made a big ass datset and trained it all together.

    You can use this lora for help other concepts to get better results, being honest i havent tested with other creators lora, just mine, but it works awesome, you can make a man thrust his meat into a brick wall,, check the gallery.

    The workflow im using is:

    *you have it inside the my video's metadata, as allways.

    Feel free to tinker and share your videos! 📹📼🎞️

    Have fun! 😜

    ***** This LoRA is provided as is and intended for lawful and responsible use.

    It acts as a general-purpose refinement layer, influencing how the base model expresses prompts and concepts rather than performing direct transformations. The resulting output is always driven by the prompt and the capabilities already present in the underlying base model.

    The LoRA does not rely on paired inputs, before/after examples, or transformation-specific training. It focuses on improving coherence, detail, and stability across generations, including mature-themed outputs already supported by the base model.

    No specialized datasets targeting identity exposure or content reconstruction were used during training. All faces and potentially identifiable features were masked or anonymized throughout the training process.

    By downloading or using this LoRA, you accept full responsibility for its use and agree to comply with all applicable laws and platform rules.*****

    Description

    FAQ

    Comments (31)

    freddiehdxdJan 8, 2026· 2 reactions
    CivitAI

    using same image you used
    a naked woman looks at the viewer for a moment, smirking. she is having sex with her leg aside, she is barefoot. sid3l3g_transition His thrusts are deep and rhythmic, creating ripples in her ass. He is slamming his hips into her ass with visible force and consistent speed. Large visible naked breasts are bouncing. When his large penis is fully inserted, she bends over in pleasure, female orgasm. 4k UHD. Static camera, fixed viewpoint, still shot.


    animation did not happen

    AbsoluteRealityJan 12, 2026· 3 reactions

    sad for you :(

    Dracken1986Feb 23, 2026

    Same, trying whatever I can, and just bounce at the moment, odd. Will keep trying :)

    PlasmadoseJan 11, 2026· 5 reactions
    CivitAI

    Can I delete the barely working Blink loras now?

    CharlieBrown0115Jan 12, 2026· 4 reactions
    CivitAI

    cant get the same results! would be cool if you just cant make some blinking loras

    ikrall001893Jan 12, 2026· 1 reaction
    CivitAI

    thanks for sharing your amazing work ! :D

    bowiba1265909Jan 12, 2026· 2 reactions
    CivitAI

    Wow! To be honest I did not expect this to work so well, but it does. Thanks for sharing. <3

    lanceshockerJan 12, 2026
    CivitAI

    What strength do you recommend to start off with?

    K3NK
    Author
    Jan 12, 2026

    you can use it standalone at 1.0 high and low, when combining i recomend like 0.4 to 0.6

    KiefstormJan 12, 2026
    CivitAI

    Thanks for the upload. By "all in one" do you mean all of your other loras in one? Do we use the prompts / triggers from those? And what does the trigger "k3nk4llin0n3" do?

    K3NK
    Author
    Jan 12, 2026

    k3nk4llin0n3 is for all the clips i used that werent the other datasets.. captions and triggerwords are untouched, so yeah, you can use those old triggerwords.. i personally find no use for this kind of triggerwording here in wan.. maybe in such a big ass dataset it makes more sense..

    machiningsolutions396Jan 12, 2026
    CivitAI

    Outstanding!! Thanks!

    g3ni3310Jan 13, 2026· 2 reactions
    CivitAI

    Love your workflow also for SFW content, quality is very high.

    Crossing my fingers for SVI integration in a upcoming version. Keep up the good work!

    K3NK
    Author
    Jan 13, 2026

    Yeah is almost ready, I managed to save latent instead of PNG, check the new version of my node in GitHub, hopefully I upload v2.5 today

    g3ni3310Jan 13, 2026

    That sounds super interesting, great idea to save the lossless latent files. Checking out your github right now.

    lolbleach001584Jan 15, 2026
    CivitAI

    hi what chekpoint or gguf you use?

    K3NK
    Author
    Jan 16, 2026· 1 reaction

    I'm using i2v 14b base comfy model

    clzpetn804Jan 25, 2026
    CivitAI

    So far I absolutely love this LoRA (as I am sure you can tell by my recent tear of generations with it), I think it actually works better if you don't use the trigger word and just make sure to be descriptive of what you want. My only non-trivial criticism is that the LoRA seems to hamper camera motion rather drastically. I have done a number of tests and in very few instances was I able to achieve the kind of sweeping camera moves I normally use with the LoRA active, but this isn't something I need for every single video and the good FAR outweighs the bad. Great stuff as usual!

    Edit: for further clarification the camera seems to move far more if I force a pure priors video by blurring the input image enough to force the model to rely entirely on priors from the LoRA and model and not forcing it to apply them to the starting image.

    oldthrashbarJan 25, 2026
    CivitAI

    General rule I find with k3nk stuff is to start the lora as low as possible because concepts tend to bleedthrough a lot. Good luck

    lakshmideesse716Jan 25, 2026· 1 reaction
    CivitAI

    BRO YOU RE AWSOME !!!!!!!!!!!!!!!!!!

    bapybaraJan 29, 2026
    CivitAI

    can someone please tell me how i can cut to a new scene in i2v like they do on pornmaker.ai

    PlasmadoseJan 31, 2026

    You use this Lora along with the words "Jumpcuts to" in your prompt. You can also google for "Blink loras".

    Dracken1986Feb 11, 2026

    @Plasmadose Do you need to include a second image as the reference?

    PlasmadoseFeb 11, 2026· 1 reaction

    @Dracken1986 No. You really don't even need this lora to do a jumpcut. Here is an example: "A woman is smiling at the viewer. The video then jumpcuts to the same woman with the same body and same clothes from the beginning of the video performing a blowjob." Be careful though, jump cutting can vastly change the appearance of the woman.

    Dracken1986Feb 12, 2026

    @Plasmadose thank you I will give it a shot now and give an update :D, been having massive issues with (at 0 seconds: ... ) and camera controls. They dont seem to be doing anything. Read something about a camera node.

    Update, yeah still not working, just stays in the same pose / animation. Would the workflow affect this?

    PlasmadoseFeb 12, 2026

    @Dracken1986 Workflow has nothing to do with it but the wan model and loras you are using does. I use Smoothmix ggufs and they work for me with most cum loras. Jumpcuts, through my own experience, aren't consistent and puts the subjects in different or weird positions your prompt may not be asking for and I think it depends on how much of the subject is in the original image. If there is already a penis in the image you are trying to generate, the jumpcut is less likely to happen. Thats why when you see most blink videos, it starts off with a near full body shot of a woman by herself.

    Dracken1986Feb 14, 2026

    @Plasmadose Thank you, thats helped alot.

    dumpy50222Feb 2, 2026
    CivitAI

    Honestly not sure how I'm getting such different results than the creator and some of the examples. I'm using the base model 14Bs and zero other LoRAs. If I apply this LoRA at any weight, the actor(s) in the scene just begin bouncing like absolute maniacs as soon as they move an inch.

    Don't get me wrong, it's funny, and there are certainly physics involved, but I'm not sure what I can do with that, or what I'm doing wrong. I've even downloaded the creator's flow and used it, same result. Pretty weird.

    Deess781Feb 9, 2026

    res_2s ? ^^

    vilhc1207671Feb 4, 2026· 1 reaction
    CivitAI

    Amazing!

    I've tried NSFW 20+ LoRA, but only this works perfectly. Thank you so much for making this!

    Dracken1986Feb 16, 2026· 2 reactions
    CivitAI

    Thank you for all your work, they are amazing. Though is anyone else getting the issue where a penis is missing the helmet and just looks like a stump? :D

    LORA
    Wan Video 2.2 I2V-A14B
    by K3NK
    Download (Beta) View on CivitAI
    deepthroat
    action
    cunnilingus
    fullnelson
    anal sex
    titfuck
    bbc deepthroat
    bbc blowjob
    sideview blowjob
    nsfw
    blowjob
    pov blowjob
    nsfw helper

    Details

    Downloads
    8,937
    Platform
    CivitAI
    Platform Status
    Available
    Created
    12/27/2025
    Updated
    5/20/2026
    Deleted
    -

    Files

    wan22-k3nk4llinon3-16epoc-full-low-k3nk.safetensors

    Size:
    292.59 MB
    SHA256:
    942a3adea87f0efbea38245357b70727c52b14f7054566080b2cdba2d218695e

    Mirrors

    HuggingFace (8 mirrors)
    CivitAI (2 mirrors)

    wan22-k3nk4llinon3-16epoc-full-low-k3nk.safetensors

    Size:
    292.59 MB
    SHA256:
    942a3adea87f0efbea38245357b70727c52b14f7054566080b2cdba2d218695e

    Mirrors

    HuggingFace (8 mirrors)
    CivitAI (2 mirrors)

    wan22-k3nk4llinon3-18epoc-full-low-k3nk.safetensors

    Size:
    292.59 MB
    SHA256:
    67b933a6aa485da9632b1e540411ed53ff451e554d42aee8aadad815598395ff

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)

    wan22-k3nk4llinon3-15epoc-full-low-k3nk.safetensors

    Size:
    292.59 MB
    SHA256:
    72bbd4b7f72d357ba17dd56b07a546b20258ba8b07c5e12ea6edd2c887056d21

    Mirrors

    HuggingFace (50 mirrors)
    CivitAI (1 mirrors)
    TensorArt (1 mirrors)
    TensorHub (1 mirrors)

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.

    SeaArt
    K3NK 4llinOne - Helper - Wan 2.2 Video LoRa - K3NK - Low noise - v1.0