CivArchive

    I created this LoRA by referring to This.
    I would like to pay my respects to the creator of that LoRA.

    (2026.4.27) Uploaded an LTX 2.3 version LoRA.
    The training techniques previously used with WAN were carried over directly.
    Since this version is a somewhat experimental LoRA, please refer to my sample video in ComfyUI and test it in various ways.
    (This has not been tested in T2V. An NSFW LoRA maybe needed.)
    Again, this is an “experimental” version. Since the step count is low, the results may not be very good. Various tests are currently being conducted, and it may be updated after retraining.

    First of all, I did not train on any facial data. As far as I know, there are already enough (or many) facial-related LoRAs (such as cum on face, cum_in_mouth, etc.).

    If want facial content, please use another great LoRA or use this one together with it (I have not tested the results when using them together).


    -- WAN --
    Using the TripleKsampler node gives even better results. I highly recommend trying it out.
    https://github.com/VraethrDalkr/ComfyUI-TripleKSampler

    Prompt used:
    She moves hand rapidly up and down the shaft of her penis and she ejaculates a thick and huge amount white sticky liquid into air repeatedly from her penis.


    I used 20 videos of 80 frames / 5 seconds each as the dataset.

    Regarding liquid ejection, compared to Wan 2.1 the results are different, and it seems to have become more difficult to generate. However, if you use the workflow of the video I uploaded as it is, the results should not be bad.

    The noise issue in the previous High_noise has been fixed, and older versions have been removed. The noise no longer occurs, and 121-frame sequences are now supported.

    It might also be a good idea to adjust the lightX lora and This lora weights, or change the prompt, to create even better results.

    Description

    • Fixed an issue where the liquid appeared in a V shape (or split into two streams).
      (It still occasionally comes out in a V shape, but it’s much better now.)

    • Since the liquid seemed to ejaculates too quickly, added a slight delay (adding “a moment later,” to the prompt increases the success rate).

    • The shaking speed has been made slightly faster.

    • also, revised and reorganized the captions and dataset.

    FAQ

    Comments (21)

    GlowingGuardianGirlNov 5, 2025
    CivitAI

    You said 'once wan2.5 is released..'. Is it confirmed that 2.5 will be public?

    swan
    Author
    Nov 5, 2025· 1 reaction

    No, it doesn’t seem to be confirmed yet.

    clawsofdeathNov 17, 2025
    CivitAI

    A "cum on camera lens / viewer" effect would be amazing for the next version, trying to prompt that but had no success so far.

    swan
    Author
    Nov 19, 2025· 1 reaction

    My cumshot LoRA is basically aimed at the same concept as “Masturbation cumshot for WAN2.1” and since it was trained mainly on data where the shot goes into the air and falls downward, So it doesn't looks easy. When data that doesn’t match the concept increases, consistency tends to get worse, so updating it is difficult as well. I’m currently overseas, so once I get back home, I’ll check if there’s any possible method.

    MCSizeMattersNov 28, 2025
    CivitAI

    Great package, very dependable, very few cumshots from the wrong place.

    But I'm struggling with getting cum to land and stick on anything. It largely lands and disappears.

    Any suggestions?

    swan
    Author
    Nov 29, 2025· 2 reactions

    I recently discovered that using this together with LOW, https://civitai.com/models/1368557/masturbation-cumshot-for-wan-i2v-14b significantly improves the liquid effect. But do not use it with HIGH for that, For HIGH use my lora only.
    It seems the training data for liquid touching or landing somewhere is rather limited. I’ll take this into consideration for the next update.

    MCSizeMattersNov 29, 2025· 1 reaction

    @swan Brilliant! It definitely helped. I tried using that lora, but used in on both low and high at the same time. I'm still trying to understand how low and high work together. Thank you!

    xDegenerateDec 22, 2025
    CivitAI

    Not bad! But I noticed this one makes the image a bit blurrier compared to your other LoRa's

    Futanari Warning https://files.catbox.moe/10sk8o.mp4

    swan
    Author
    Dec 23, 2025

    I watched your video, Must be use both high and low together, and the results can vary depending on the workflow settings or video interpolation. If you drag and drop my sample video into Comfy, the workflow will appear, so please take a look at it.

    DoitDoitJan 17, 2026· 1 reaction
    CivitAI

    It works very well and is great, thank you bro!!

    caiamongFeb 7, 2026· 1 reaction
    CivitAI

    The nipples seem to flap around as if the breasts are big and bouncy even when they're practically flat chested. Otherwise the cum effect is great.

    swan
    Author
    Mar 7, 2026· 1 reaction

    Issue confirmed. The fix will be included in the next update.

    Nk0GGryMar 2, 2026· 1 reaction
    CivitAI

    this is a great lora. but how do I get the white liquid to go in any direction but outward in an arc away from the subject?? like for example, toward the subject's own chest area...

    swan
    Author
    Mar 7, 2026· 1 reaction

    There isn’t much training data showing the liquid hitting or landing on something, so it may be difficult to implement properly. This lora was trained mainly on liquid spilling/flowing outward. Since I’ve received a lot of feedback about liquid landing on or touching surfaces, I’ll test it in the next update, but if the landing position isn’t consistent, the difficulty increases, so I’m not sure how well it will work.

    MCSizeMattersApr 24, 2026
    CivitAI

    Any thoughts on applying your skills to LTX 2.3? It could really use a cum lora, there it literally nothing.

    swan
    Author
    Apr 25, 2026· 1 reaction

    I just tried LTX 2.3, and it seems I underestimated this model without even using it first.

    This is interesting... it looks like a very promising model.

    Since the tool I use supports LTX 2.3 training, making LoRAs for it doesn’t seem to be a problem.

    However, it still seems like more testing is needed for now.

    News will come soon.

    swan
    Author
    Apr 25, 2026

    It seems the training tool I’m currently using doesn’t support LTX 2.3 yet.
    I’ll either have to find another method or wait a little until the training tool supports it.

    MCSizeMattersApr 25, 2026

    @swan Totally get that. LTX 2.3 is incredibly impressive, lightyears ahead of Wan in so many ways, especially adding sound. But just like z-image at the start, it lacks the needed adult, um, plumbing so many of us use.

    swan
    Author
    Apr 27, 2026

    @MCSizeMatters But well, it’s done now.

    MCSizeMattersApr 27, 2026

    @swan Amazing! I just downloaded it and will give it a try, I'll post results!

    MCSizeMattersApr 27, 2026· 1 reaction

    @swan Nice job, you unblocked the cumshot in LTX 2.3, your lora works really well right out of the box and replaces a hodgepodge of several other loras I was trying to use.

    You unblocked the cum for LTX 2.3!