CivArchive
    Discovering a morning wood [Real] | Wan 2.2 I2V/T2V - v1.0 [high]
    NSFW

    ⚠ My LoRAs will be posted on HuggingFace from now on ⚠

    Omg, it's a morning wood?! 😱

    Yeah stepsis, it is 😔

    Civitai Censoring

    This LoRA will probably be deleted (which is pretty stupid, since Wan 2.2 without any LoRA is perfectly capable of removing someone's clothes).

    Other amazing LoRAs have been deleted, and this one will probably too.

    To see all my LoRAs, visit my repository on HuggingFace. There you'll find examples, prompts, instructions, and download links.

    See all my LoRAs on HuggingFace

    Training

    • 200 steps on high noise

    • 1 edited video (81 frames)

    • Lots of hope

    Usage (Low noise)

    This LoRA was trained only on High Noise, which means you'll have to use some Low Noise that knows what a penis is

    (I used the Low Noise from DR34ML4Y v2, which is still available on Civitai).


    Got an idea for LoRA? Then let's do it! (or at least try 🤠)

    In my repository, I created a model called "LoRA Requests", so if you want to request a LoRA, send videos and photos to help build the dataset for a public LoRA that will be posted here, that's the place to go.

    Open LoRA Requests on HuggingFace 🤗


    I2V:

    Prompt:

    mrn_w00d, The scene starts with a woman, then the scene cuts to bedroom indoors, the same woman enters through the opened door, entering in the bedroom.
    
    She approaches the camera, opens her mouth in surprise and lift up the duvet, revealing a man's hard penis and thighs that was underneath.
    
    The scene cuts to the same woman in closeup, swallowing the man's hard penis. With both hands she holds the base of the man's hard penis and slowly lifts her head until the man's hard penis is completely out of her mouth. Lines of saliva connect her lips to the sticky hard penis.

    High Noise LoRA Scale: 1.5

    Low Noise LoRA Scale: 1.0

    Shift: 4

    T2V:

    mrn_w00d, The scene starts with a latina woman wearing sexy pajamas, in a bedroom indoors, she enters through the opened door, entering in the bedroom.
    
    She approaches the camera, opens her mouth in surprise and lift up the duvet, revealing a man's hard penis and thighs that was underneath.
    
    The scene cuts to the same woman in closeup, swallowing the man's hard penis. With both hands she holds the base of the man's hard penis and slowly lifts her head until the man's hard penis is completely out of her mouth. Lines of saliva connect her lips to the sticky hard penis.

    (Same as I2V)

    Help me creating more

    If you want to help me continue making LoRAs, or if you want me to make a LoRA for you, buy 5000 PlayCoins at Aisha-AI and transfer them to my account (account number 2).

    This helps Aisha-AI to stay alive and produce new LoRAs for you all 💜

    Description

    First version

    FAQ

    Comments (13)

    boobkake22Feb 23, 2026· 4 reactions
    CivitAI

    I suspect this will get taken down, but it's a fun concept! The Low Noise file appears to be missing?

    AishaAI
    Author
    Feb 23, 2026· 4 reactions

    Just read the post

    RosettaseesFeb 23, 2026
    CivitAI

    Works pretty well. Though did experiment with two people in the start image and it tended to lose face consistency on one of the people when generating.

    AishaAI
    Author
    Feb 23, 2026

    Well, that was totally expected since I did the training with only 1 video and only 1 person. Even so, it's possible to scale down to 1 or less, but you'll need to search extensively for a prompt that can maintain consistency.

    RosettaseesFeb 23, 2026

    @AishaAI Oh totally, that wasn't a complaint, just an observation. As said it works well. Looking forward to more of your work.

    AishaAI
    Author
    Feb 23, 2026· 2 reactions

    @Rosettasees I know it's not. This LoRA practically overfitted, it only didn't turn out bad because it's only the High Noise. But it's possible to achieve other things by lowering the Scale and changing the prompt... But with a high Scale, it will always force this scene progression with one person, which was the idea of ​​LoRA.

    I'm also eager to see more of my own work 🤣🤣🤣 The problem is that I spend a lot on training that doesn't work, in addition to the normal expenses that are necessary. Training only the High Noise saves me money, but I still end up wasting a lot on attempts.

    NeoxenaFeb 23, 2026
    CivitAI

    Very Nice!, I missed the Bath one so I'm sad but this one works well.

    AishaAI
    Author
    Feb 23, 2026· 2 reactions

    The Bath LoRA was already blocked by Civitai, but that LoRA wasn't really good, it was just functional. Maybe I'll make another one and post it, this time as T2V so it doesn't get removed 😁

    BRai4nowFeb 26, 2026· 2 reactions
    CivitAI

    It's pretty insane how well this works. I wish everyone's loras tried to maintain the look of the inference image as well as this lora.

    AishaAI
    Author
    Feb 26, 2026· 4 reactions
    CivitAI

    Civitai is 💩
    I'm posting my LoRAs on HuggingFace

    TieFighterPilotFeb 27, 2026
    CivitAI

    Which checkpoint did you use? Because the scene cuts may probably will not work correctly, I have not tested them but it normally happen, like it gets overrided, perhaps thats why the strenght is above 1.0 right?

    Edit: Your work is great! I dont know why your LoRAs get deleted, any LoRA can be achieved with just good prompting 🤷🏻‍♂️

    AishaAI
    Author
    Feb 27, 2026

    Scene cuts are embedded in LoRA, so it should work at any checkpoint.

    But the LoRA training was done entirely on Wan 2.2 Base (wan2.2-t2v-a14b-diffusers)

    TieFighterPilotFeb 27, 2026· 1 reaction

    @AishaAI Yes it works perfectly, It works with euler-simple, I was using lcm-normal doesnt work with that, thanks a lot this lora is fun!