CivArchive
    Preview 81136292
    Preview 81132053
    Preview 81132668
    Preview 81132534
    Preview 81132885
    Preview 81133008
    Preview 81132148
    Preview 81133148
    Preview 81132300
    Preview 81132591
    Preview 81132337
    Preview 81132756
    Preview 81132973
    Preview 81133473
    Preview 81134023
    Preview 81133676
    Preview 81133756
    Preview 81133791
    Preview 81135981
    Preview 81135251

    Unlock xxx images using flux, Qwen or Wan 2.2.

    ATTENTION

    Second version for WAN 2.2 T2V works with the usual models BUT for T2V I recommend some different models and LoRas:

    1. https://huggingface.co/lightx2v/Wan2.2-Lightning/blob/main/Wan2.2-T2V-A14B-4steps-250928-dyno/Wan2.2-T2V-A14B-4steps-250928-dyno-high-lightx2v.safetensors instead of wan 2.2 high

    2. https://huggingface.co/eddy1111111/WAN22.XX_Palingenesis/blob/main/WAN22.XX_Palingenesis_low_t2v.safetensors instead of wan 2.2 low Removed this recommendation

    3. NO speed/low step loras on high noise phase

    4. https://huggingface.co/alibaba-pai/Wan2.2-Fun-Reward-LoRAs/blob/main/Wan2.2-Fun-A14B-InP-low-noise-HPS2.1.safetensors with STR:1.0 on low noise phase

    5. https://huggingface.co/lightx2v/Wan2.2-Lightning/blob/main/Wan2.2-T2V-A14B-4steps-lora-250928/low_noise_model.safetensors with STR:1.0 on low noise phase

    6. ???

    7. Profit!

    All my videos have workflow embedded, the workflow is just a suggestion. But it works, and works very well.


    First version for KREA.

    Start using: 20 steps, res_2s, bong_tangent, STR 0.8

    After that you can experiment with other arguments. It is not very good and deform hands and feet very badly.


    ALL PICTURES ARE PNG WITH WORKFLOW INCLUDED


    It is not perfect, sometimes it creates monsters, detached members and strange things, just try another seed or adjust your prompt. Eventually it get it right.

    Examples:


    Recommended Usage:

    • Checkpoint: flux1-dev

    • Strength: 1 (adjust depending on the image)

    • Sampling Method: DPMPP_2M

    • Scheduler Type: BETA

    • CFG Scale: 4

    • Steps: 35


    Feel free to share your creations and feedback so I can improve this model and continue contributing to the creative community!

    Description

    It is a small step in the right direction for sure.
    Improved anatomy quality in general. Included the hentai concept. Less horrors on this version.
    BUT IT IS NOT PERFECT.
    Sometimes you will need to adjust your prompt, try other seed or add another LoRa to get the desired effect.

    FAQ

    Comments (15)

    puma0000Jun 9, 2025· 1 reaction
    CivitAI

    Try as I might, I can't get photorealistic style images, it defaults to anime, even with descriptive prompts. Is there a different weight or CFG I should be using?

    alcaitiff
    Author
    Jun 9, 2025

    What prompt is causing issues to you? Let me try it.

    alcaitiff
    Author
    Jun 9, 2025

    Anyway, to guarantee a style I recommend you to use a style LoRa. Look at this 2 images:
    https://civitai.com/images/81305060
    https://civitai.com/images/81305221
    One with realism LoRa and one with anime LoRa.

    If you want me to help with any prompt just let me know.

    flo11ok874Jun 9, 2025· 3 reactions

    Maybe try different checkpoint if you use flux.dev? Real Dream or Fluxmania are greater then flux. And work fine with this mod.

    alcaitiff
    Author
    Jun 23, 2025

    Try v6. It should be better now.

    flo11ok874Jun 23, 2025· 1 reaction

    @alcaitiff I don't use flux dev but v6 broke photos with RealDream and FLuxmania when v5 and v4 works great with them. Just for you know if someone ask, still v4 is very cool so I'm still stay with it and thanks for that.

    alcaitiff
    Author
    Jun 23, 2025· 2 reactions

    @flo11ok874 v6 improved a lot of things, but custom lora and checkpoints appears to not like it. I'll work on that.

    Soup_FusionJun 10, 2025· 2 reactions
    CivitAI

    overall this is probably the best generalized NSFW lora available. that said i have noticed that it will default to an anime style here and there, and it tends to create pretty similar looking women, and sometimes the women look a little too young no matter the prompt, for some reason this especially prevalent when creating cumshot images.

    alcaitiff
    Author
    Jun 10, 2025· 4 reactions

    Yeah, I was under the impression that people would prefer a generic LoRa. That could be used with any other LoRa. And for this I need mixed styles in the training. But a lot o comments seems to want a fixed style. So I will create a Hentai and Realistic only versions. To ensure people get what they want.
    About the young, It may appear to be the case because cartoon images translate bad to adult woman. It's may be hard to the model understand. This may word better on the next version.

    KriptofaustJun 19, 2025

    @alcaitiff This news is very encouraging. can you give us an estimated release date?

    alcaitiff
    Author
    Jun 19, 2025

    @Kriptofaust the hentai and cartoon versions are already out. I'm working on the realistic one. If you have any hard prompts that you want me to improve please send me by dm

    alcaitiff
    Author
    Jun 19, 2025· 1 reaction

    @Kriptofaust About the date, it's hard to estimate. It will depends of the model wiliness to cooperate, sometimes it's fast 1 or 2 days. Sometimes it take a month.

    alcaitiff
    Author
    Jun 23, 2025

    OK, v6 is out. Give it a try.

    joshlebitz2968Jun 23, 2025· 1 reaction

    @alcaitiff You have probably one of the best NSFW LoRAs for Flux, it's actually amazing.

    Thank you for your hard work. Is it possible to add this position on the next update? https://civitai.com/models/560154/straddling-handjob

    alcaitiff
    Author
    Jun 24, 2025· 1 reaction

    @joshlebitz2968 it will be hard to find good pics of real people doing that to make a dataset. But I can try