With this Lora, you can peek the scene of your Waifu being cuckolded from the ceiling camera. I have had Lora learn several sex positions, but the results generated are somewhat erratic. Please refer to the Recommended Prompts and Sex Position Prompts.
Strength: around 0.8
CFG Scale: around 5.2
Triger Word:
NTRmonitoringSPRecommended Prompts:
(from above:1.8), from side, bird eye view, bird's eye view, depth of field, couple focus, 1boy and 1girl, vignetting, darkness, dark room, dark environment, full body, barefoot, thighs, arms, scanlines, vhs video, messy room, floor, window, curtains, bookshelf, potted plant, BREAK 1boy, ugly male, fat male, dark skin male, faceless male, naked male
BREAK 1girl
viewfinder, recording, battery indicatorSex Position Prompts:
standing sex from behind, bent over, living room, table, arched back, grab torso, barefootkiss, kneeling on bedmissionary, (pillow:1.3), (large bed:1.2), vaginal sex, spread legs, on bed, feet, grab thighsgirl on top, cowgirl position, (pillow:1.3), (large bed:1.2), vaginal sex, on bed, thighs, barefoot, feet, cowgirl positiondoggystyle sex, (pillow:1.3), (large bed:1.2), on bed, grab torso, ass, all fours, feet, shoulders(after sex:1.2), cum on body, cum in pussy, cum overflow, fingering, spread legs, lying on back, breath, open mouth, completely nude, stained sheet, male kneelingDescription
FAQ
Comments (8)
Beautiful 😍😍😍
😘😘😘
Love this Lora, trying to make something at the moment but struggling. Do you have any advice on how to keep the background room the same (so different shots in the same room for different videos) as well as the battery, Ive tried a few prompts but the battery always seems to bounce around, trying to get it to stick to a specific %. Thanks again.
Thanks a lot, glad you like the LoRA!
For keeping the background consistent, the simplest way is to repeat the same key objects in your prompt (e.g. “bookshelf, potted plant, window, blue curtains”) and keep the camera angle similar. You can also use img2img or a reference image to better “lock” the room while still allowing variation in the subject.
For the battery, that’s a bit harder since models don’t handle exact numbers very well. A few things that can help:
･ Describe it visually, not just numerically (e.g. “battery icon half full, 50%, clearly half-filled”)
･ Reuse a good result with img2img to keep it consistent
･ Inpaint the battery area
In general, backgrounds are much easier to stabilize than small UI details like a battery indicator, so combining prompt consistency + image-based methods tends to give the best results.
@MNeco Hi, thank you very much for your advice, my issue is I am so used to Wan2.2 where you just write the prompt in sentences. Still trying like mad to get you to this method and the structure (did find: https://danbooru.donmai.us/wiki_pages/tag_groups helpful though). I have yet to find a Img 2 img workflow that ... embarrassingly yet honestly... I understand, do you have any suggestions and I will have a look.
Again thank you very much for your work, time and also your reply / advice, really appreciated.
@Dracken1986 Thanks again, really appreciate your kind words!
Yeah, moving from sentence-style prompting (like Wan2.2) to tag-based prompting can feel a bit unintuitive at first. A simple way to think about it is: instead of writing a sentence, you’re just listing the most important visual elements clearly and consistently.
For example, instead of:
“a person sitting in a cozy room with a bookshelf and window”
you’d write something like:
“cozy room, bookshelf, window, indoor lighting, [subject details]”
Keeping those “anchor” elements identical across prompts helps a lot with consistency.
If you’re using ComfyUI (guessing from “workflow”), I’d suggest starting with a very simple img2img setup:
Load Image → KSampler → VAE Decode
Set denoise low (around 0.3–0.5)
Keep the environment tags the same, and only change the subject
That’s honestly enough to get started — no need for a complicated workflow.
If img2img still feels confusing, you can ignore it for now and just focus on consistent tags first. That alone already helps a lot, and you can layer img2img on top later.
Once it clicks, it becomes much more predictable 🙂
@MNeco Cheers again, thanks to your advice, managed to make the images used for a quick NTR video that kept the background rather similar (Much better than what I was generating).
@Dracken1986 Wow, you have amazing skills. That’s wonderful!
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.