Porn Factory AILogin

AI porn models for adult image and video generation

Browse the building blocks behind the factory. Filter checkpoints and LoRAs by image or video support, base model, and tags, then open any version directly in the unified generator.

1487 models available

LoRA · SD

Jinx League of legends

30
This LoRA model, trained on Jinx League of Legends images, can reproduce the character convincingly with outfit flexibility and accurate tattoo depiction. The trigger word is JinxLol. It performs well with NSFW images.characterdigital illustrationgame characterillustrationjinx league of legendsleague of legendsvideo game

LoRA · SD

Tifa_meenow

50
Tifa_meenow is a Lora model for generating images of Tifa from Final Fantasy VII Remake, trained with a mix of real people and game assets. It aims for flexible use with other Loras but may require tag adjustments to avoid nudity. The model excels at facial modeling and is recommended to be used with a weight of around 0.75 for v2.final fantasyfinal fantasy vii remaketifatifa lockhart

LoRA · SDXL

FF-Minecraft-XL (+Dataset)

00
FF-Minecraft-XL is a model that generates images in the style of Minecraft. It utilizes custom SDXL models and is based on pre-rendered, AI-generated Minecraft images. The model is available in various Lora formats.animedatasetminecraftstylewoman

LoRA · SD

Raiden Shogun | Realistic Genshin LORA

50
This LoRA model generates realistic images of Raiden Shogun from Genshin Impact, trained without using real person's faces. It's designed to be used with the additional-network webui plugin, with recommended weights between 0.55 and 0.75 for optimal results.animecharactergame charactergenshin impactphotorealisticraiden shogunvideo game

LoRA · Wan Video 14B t2v

CGI Disney Princesses for Wan 14B T2V+Hunyuan T2V

10
This model is a Hunyuan video lora for generating CGI-style videos of Disney Princesses based on "Ralph Breaks the Internet" designs. It allows for creating likenesses of various princesses and even custom characters, with the strength influencing the level of CGI and cartoon effects. It supports multiple princesses in a single prompt and is compatible with NSFW content.annaarielaurorabellecgicharactercinderelladisneyelsaprincessrapunzelsnow whitevideo

LoRA · SD

concept Female ejaculation

60
This model enhances the visual representation of female ejaculation in images. It works best in frontal views and may require adjustments to the prompt weight depending on the base model used to balance effect strength and avoid distortions. Recommended prompts include "female ejaculation," "female orgasm," "pussy juice," and "ahegao."animeconceptejaculationorgasmpussy juice

LoRA · SD

blindbox/大概是盲盒

140
animeartstylechibicutestyle

LoRA · Other

St. Louis (Luxurious Wheels) (Azur Lane)

130
Sexiest dress on the sexiest girl! (showcase images used kl-f8-anime2 vae)animeazur lanecharactergame characterhentaist. louisvideo game

LoRA · SD

Body Horror Creatures

20
This LoRA generates disgusting body horror creatures, inspired by movies like "The Thing." It works best with the prompt "disgusting body horror creature" and can be used with different base models for realistic or anime styles. Lower LoRA levels can create nice creature effects, while higher levels produce more extreme results.body hororcreaturesdisgustinggoregrosshorrormonstersstyle

LoRA · Wan Video 2.2 T2V-A14B

Instagirl WAN 2.2

10
Instagirl v2.5 is a foundational style LoRA designed to create a realistic aesthetic, intended as a base layer for adding character or concept LoRAs. It's engineered for diversity and realism, with included high-noise and low-noise variants, and can be used for creative and commercial projects, but not for resale or hosting on competing generation services. The model requires the trigger word "Instagirl" and has a recommended strength of 1.0.hyperrealisminstagramstylewan2.2

LoRA · SD

See-Through Dress

20
This model generates images of subjects wearing see-through dresses. It's trained on a dataset of 40 images with specific settings outlined for optimal use, including sampler, CFG scale, and LORA weight. Key prompts include "wearing see trough dress," "elegant dress," "lingerie," and "back view."clothingfemalegirlsphotorealisticrealisticsexywoman

LoRA · SD

Pant Pull Down - LoRA

60
The Pant Pull Down LoRA model is designed to generate images of characters with their pants pulled down. It is trained on approximately 100 images and works best with prompts including rear view angles and adjustments to the LoRA weight. It may exhibit issues with faces and hands, but these can be mitigated with prompt adjustments.bootypantiesphotorealisticstylewoman

LoRA · Wan Video 14B t2v

SideView Deepthroat T2V Wan2.2 Video LoRa - K3nk

10
This LoRa model, trained on a dataset of video clips, focuses on generating images of deepthroat acts. It aims to accurately depict the penis's anatomy and comes in two versions optimized for different noise levels and anatomical detail. The creator encourages experimentation and responsible use.actionblowjobdeepthroatnsfw

LoRA · Other

Photorealistic Succubus Womb Tattoos (Flux, SDXL, SD1.5) 魅魔逼真私密纹身

30
This LoRA generates photorealistic tattoos located between the navel and vagina. It is compatible with SDXL and SD1.5, and subsequent versions offer improved fidelity, clearer lines, consistent shapes, and better compatibility with other LoRAs. The trigger word is "tattoo on stomach."art styleclothingcrotch_tattoofantasyphotorealisticpubic tattoosexysuccubustattooswoman

LoRA · SD

Niji二次元风格/Niji V5 style Lora

00
This Lora model creates images in the Niji anime style, encompassing Default, Expressive, and Scenic styles. It is designed to work with flexible base models like Ghost to achieve diverse outputs from a single Lora, challenging the idea that each Lora is limited to one style. Recommended usage involves adjusting the "realistic/scenery/pixiv" tag in prompts and using a weight between 0.8-1.0.animebackgroundsboysfemalegirlsmidjourney stylemidjourney v5nijinijiv5scenerystyle

LoRA · SD

Doggystyle from side view

70
This LoRA generates images depicting the doggystyle sexual position from a side view, trained primarily on Anythingv4.5. It aims to improve image quality and reduce anatomical errors, though some issues with limb generation and prompt complexity remain. It's suggested to use a weight of 0.6-0.8.2danimehentaipornstyle

LoRA · SDXL

Pixel Art XL

50
Pixel Art XL is a model designed for generating pixel art images. It achieves optimal results when downscaled eight times using Nearest Neighbors and with a fixed VAE to prevent artifacts. No specific trigger words or style prompts are required for usage.pixel artpixelartstyle

LoRA · SD

Vivid Impactful Style (Yoneyama Mai [米山 舞] Style Likeness) - LoRA/LoCon

40
This LoCon model, Vivid Impactful Style, generates images in the style of Yoneyama Mai. Use the trigger word "vvi(artstyle)" for the general style, or "vvi(face)" and "vvi(eyes)" for portraits, and "vvi(oa)" for official art-like outputs, with a recommended weight of 0.6-1. It is compatible with models like Anything V4.5 and Dan Mix.animeart styleartstyleillustrationstyleyoneyama maiyoneyama mai style

LoRA · SD

Haute Couture | Pencil Dresses

20
This model generates images of pencil dresses, a staple in everyday and office fashion. It allows users to create images of sharply dressed figures in pencil dresses. Consult prompts for secondary trigger words.dressfashionfashion designwoman

LoRA · SD

koreanDollLikeness_v10

30

LoRA · Wan Video 14B i2v 720p

Live Wallpaper Style

10
This LoRA reproduces a live wallpaper video style, similar to League of Legends launcher videos, and can be used for lofi videos. It was trained on curated video datasets with varying parameters and includes several versions optimized for different models and features, often incorporating additional LoRAs for enhanced motion, cinematic depth, and temporal control. A trigger word, when applicable, is needed in prompts.anineconcepthunyuanlive wallpaperlofilolltxvwallpaperwan

LoRA · SD

小人书·连环画 xiaorenshu

90
This LOAR model, 小人书·连环画 (xiaorenshu), commemorates the style of Chinese comic books, also known as "Little People's Books," which were a popular art form in China. The model aims to recreate this increasingly distant art style and users can experiment with CFG values and weights for varied effects. The model will be continually updated with various stylistic versions.2dart styleartstyleillustrationmonochromestyle

LoRA · SD

JK uniform

40
This model generates JK uniform clothing items. It includes various colors and styles of shirts, skirts, ties, suits, and sailor-style tops.clothingjkjk stylejk uniform

LoRA · Other

Feet Pose (anime)

50
This model generates anime-style images focusing on feet, particularly soles, in a sitting pose. It is designed for use with anime models, and the creator recommends specific prompt structures and LORA weights for optimal results, while acknowledging potential AI limitations with realistic foot generation.animebarefootfeetfemalefoot fetishposes
Load more models

Showing 24 of 1487

How checkpoints and LoRAs fit together

Start with a checkpoint

A checkpoint controls the base rendering system. Its supported modes determine whether it can create images, edit an input, or produce video.

Match the base model

A LoRA must target the same base-model family as the checkpoint. Similar names do not guarantee compatible weights.

Check the validated workflows

A catalog version can be documented before it is generator-ready. Only versions with validated broker and artifact mappings can spend credits.