2025/08/15 update
This version is still in beta, so you may encounter some unnatural results. If you're not satisfied, please keep generating until you find a result you like.
I've trained multiple versions with the help of experienced developers, and this one represents a significant step forward. It's built on a solid foundation that I know how to improve in the future.
However, the key training steps and parameters, especially for I2V (Image-to-Video), don't have a perfect "magic number" yet. Once I discover a more effective way to enhance this model, I'll release an update.
I know this isn't the final or perfect version, but I'm very aware of that fact. So, for now, please enjoy experimenting with it!
🎯 Anime Cumshot Aesthetics – Precision Load
Anime Cumshot Aesthetics – Precision Load is a LoRA designed to enhance the density, precision, impact, and aesthetic realism of cumshots in anime-style scenes.
It focuses on thickness, directional accuracy, and artistic expression, delivering short yet impactful visuals. While optimized for anime, it can also pair with live-action content.
This version does not include an off-screen ejaculation scene.
The following are my observations and conclusions from my personal experiments over the past few days, and they are not necessarily correct.
The most powerful aspect of WAN 2.2 is the High Noise model, which excels at establishing the video's motion and core framework. This is where the biggest improvement lies. For this reason, I suggest setting the steps for High Noise to a minimum of 3 steps. If you prefer a simpler approach, you can set the High and Low Noise steps to the same value.
The Low Noise model is a fine-tuned version of WAN 2.1. Its purpose is to remove noise and add detail to the dynamic framework generated after the High Noise training. Based on this, I believe the High Noise training steps can be half of the Low Noise steps, or even less. For this version, I used around 2,000 training steps for High Noise and over 6,000 for Low Noise.
Regarding the Low Noise LOSS value, the behavior in WAN 2.2 is different from 2.1, so it's hard to pinpoint the ideal value. I tested each epoch and chose the version with the lowest LOSS. My source material is a mix of anime and real-life footage, so I'm not sure if it's overtrained—but it's a good place to start.
📁 Training Details
Clips: 39 original animated clips
Length: 3 seconds each
FPS: 16
Style: 50% anime, 50% live-action
Type: Emotion-focused enhancement
high noise
low noise
dataset.toml
PS: When training high noise, you can use a resolution of [192, 192].
💡 WAN2.2 has strong built-in emotional reactions. My dataset also includes extra facial expression training. Writing detailed prompts will yield the best results.
All my video templates are upscaled, smoothed, and interpolated using my V2V (video-to-video) workflow.
⚙️ Suggested Settings
Steps: 8 or 10 (High noise 4 + Low noise 4, or High noise 5 + Low noise 5)
CFG Scale:
CFG: High noise: 2.0 ~ 2.5 (usually 2.0)
CFG: Low noise: 1.0
Sampler: Euler / Euler a/ heun
Scheduler: SIMPLE/ beta/ beta
You can refer to my images, and I'll also upload my training data.
💬 Usage Notes:
Used with the official workflow (KJ workflow optional)
No acceleration in high noise; low noise uses lightx2V 256 LoRA accelerator
High noise LoRA strength: 1 (I rarely use a strength of 2, typically only when I'm testing multiple LORAs simultaneously. Otherwise, I always keep it at 1.)
Low noise LoRA strength: 1 (lowering effect unknown)
🎭 Emotional Reaction Examples
WAN2.2 supports a wide range of emotions and movements. Examples (mix and match freely):
Please use your imagination and creative writing skills to write some content based on the following emotional reactions to the new WAN 2.2 model: Surprised, Angry, Accepting, Avoiding, Happy, Sad, Fear.
The prompt is that the WAN 2.2 model is incredibly powerful and follows prompts very well. Try to develop content that reflects these emotional reactions to the model's capabilities.
Surprised
As her face is hit, she flinches, recoiling slightly, eyes widening in confusion.Angry
Her brows furrow, eyes glare, lips part slightly, exposing her teeth.Accepting
Her expression remains calm, lips slightly parted, receiving the cum directly.Avoiding
She instinctively turns her head away, causing subsequent shots to land on her hair.Happy
A radiant smile spreads across her face, eyes locking with the camera.Sad
Fear
Emotion：
Here are all my available packages for your reference. You're welcome to mix and match, but I haven't tested too many combinations for success, so please try it yourself!
I originally used this emotional version in WAN2.1.
Her face shifts in an instant. She looks shocked, then tilts her head upward, staring toward the source of the cum.
A sudden jolt crosses her face. She flinches in shock, slightly recoiling as she slightly turns her head to the right.
She reacts immediately. Her eyes widen in shock, then she turns toward the source of the cum, looking very angry.
There is an instant shift in her expression. She flinches in shock, her eyes wide with disbelief.
She blinks slowly, without changing her expression.
A wave of shock washes over her face. She flinches in shock, slightly recoiling.
A sudden emotional shift appears on her face. She looks shocked, then turns toward the source of the cum, looking very angry.
A sudden change crosses her face. She flinches in shock, her eyes wide with disbelief.
There is a subtle shift in her expression. She blinks slowly, frowning faintly, her expression calm but clearly displeased.
A moment of disbelief flashes across her face. She looks incredulous, then tilts her head upward, staring toward the source of the cum.
Her expression remains unchanged.
Her face tenses briefly. She tilts her head upward, eyes following the source of the cum.
A quiet reaction passes over her face. She closes her eyes for a moment, frowning faintly. Her lips press together tightly, her face calm yet clearly displeased.
A quick flash of shock crosses her face. She flinches in surprise, her eyes wide with disbelief.
Initially, she is momentarily stunned. Then, flinching in shock, her eyes squeeze shut, and she instinctively recoils, noticeably turning her head to the right.
Her expression changes instantly. She furrows her brow, her mouth agape in disbelief and rage. She then looks up and glares toward the source of the cum, clearly furious.
As the fluid hits, she flinches in surprise, blinking a few times, then noticeably turns her head to the right.
A sudden wave of emotion strikes her face. She flinches in shock, her eyes wide with disbelief.
Her face is flushed. She stares ahead with hazy eyes, seemingly dazed yet still conscious, her expression calm and unresisting. Her left eye had briefly closed before reopening.
Her features soften. She smiles faintly and keeps her gaze on his penis without looking away.
Her brows twitch faintly. She closes her eyes for a moment, then reopens them with a slight smile.
A flash of shock flickers across her face. She flinches, her eyes wide with disbelief.
Her expression shifts suddenly. She flinches, then turns toward the source of the cum, her eyes wide with disbelief.
Her features soften slightly. With a soft smile, she lifts her head and looks up at him, letting the cum splash across her face and chest.
A sharp reaction crosses her face. She flinches in shock, her eyes wide open, then turns toward the source of the cum.
Her expression remains unchanged, her eyes simply closing.
She shifts her head slightly and watches him quietly, expression calm but aware.
A quiet disapproval flickers across her face. She blinks slowly, frowning faintly, then raises her head, her expression calm but clearly displeased.
As the fluid hits, her eyes widen in shock, and her mouth slightly opens, as she noticeably turns her head to the left.🖋️ Prompt Building
Basic Poses
Mix and match as desired. However, success isn't guaranteed, so please try it yourself or refer to my video examples.
A girl is kneeling.
A girl is sitting.
A girl is standing.
A girl is lying down.
Entry Actions
If you want the man to not be in the frame at the beginning and then enter the frame from off-screen, use this.
A man's penis enters the frame from the bottom left corner. (right corner)
A naked man enters from the right side. (left side)
A man's penis enters from below.
Main Actions & Landing Points
Main actions and landing points
The man is stroking his penis. He ejaculates on her face / in her hair / on her chest / into her mouth / on her tongue / stomach.
PS: My LORA is primarily focused on the face, with other areas being a secondary effect. If you want it to land specifically on the hair or breasts, please use playtime_ai's "Body Cumshot (Wan 2.2 / 2.1)". My LORA can't easily target a specific location; it mainly hits the face and then splashes onto other parts.
Body Cumshot (Wan 2.2 / 2.1) - v2 Wan2.2 - T2V - LOW | Wan Video LoRA | Civitai
Assisted Actions
If you want the woman to help the man, use this. For other scenarios where the woman helps the man, like using her breasts, I apologize, but I don't have that material. However, you can try it; theoretically, the ejaculating action shouldn't disappear
She strokes his penis. He ejaculates into her mouth, with a small amount landing on her nose.
💧 Describing Cum Texture
Describe the state of the semen （Please use this as a reference to capture a similar feel）
The cum is thick and sticky, clinging like paste before sliding off.
Thick and heavy, the cum holds to her body before dripping away.
The sticky fluid briefly clings before slowly sliding off her skin.
The cum is thick and gooey, clinging to her lips and tongue before dripping off
It is thick and sticky, clinging like paste before sliding off slowly
⚠️ Current Limitations & Known Issues
Occasionally, cum may appear on the face before ejaculation starts.
In certain angles, the fluid may appear unevenly distributed or repeatedly hit the same area.
These artifacts rarely occur in the original training data but may still appear in the generated output.
Exact solution is still unknown — success rate remains around 80% based on personal testing, and already stable enough for most use cases.
🔗 Recommended Pairing
Works standalone
Can be mixed with other LoRAs (untested)
Experiment with combinations in your own workflow
🛡️ Disclaimer
All characters are AI-generated and do not represent real people.
You are solely responsible for all generated content and its use.
This LoRA is exclusive to Civitai — do not repost or upload elsewhere.
Description
2025/08/15 update
This is the extended WAN 2.2 version, based on my original WAN 2.1 "Anime Cumshot Aesthetics – Precision Load" model.
It's divided into High-Noise and Low-Noise versions, so you'll need to use both LORAs. If you're unsure how to use them, please research the basic usage of the WAN 2.2 version first.
This version is still a beta, and you may find some unnatural results. I'll train the next version when I have better source material or a solution for these issues. For now, this version is for you to experiment with.
All of my video examples were created using my own LORA models exclusively, with no other LORAs added, except for the lightx2V 256 version acceleration LORA used in the low-noise version.