Premium character LoRAs
Original synthetic characters, bundles and AI tools are in my shop:
Join the community on Discord: https://discord.gg/nB79wRNtMT
BlowJob
A concept LoRA for Krea 2 focused on POV oral sex: a woman kneeling or leaning in, lips wrapped around an erect penis, often looking up at the camera, topless, fully nude or in a bra, hoodie or tank top, with glasses, headphones, chokers and piercings. It gives an amateur photo look, shot from above, in bedrooms and on carpets and rugs.
Usage
Start with the trigger word BlowJobRK, then describe the woman (skin, hair), what she wears or that she is topless or nude, the act (lips wrapped around his erect penis, tongue, eyes looking up), the camera angle and the place, and end with the light. Saying that the camera is from above and that she looks up at it brings out the POV effect the most.
BlowJobRK, Photograph of a woman with <skin> and <hair>. She is <topless / nude / wearing ...>. She is kneeling and performing oral sex on a man, her lips wrapped around his erect penis, her eyes looking up at the camera. <Place>. The camera angle is from above. The lighting is <light>.
Recommended weight: 1. Made for Krea 2 Turbo (8 steps, CFG 1).
Prompt Examples
"BlowJobRK, Photograph of a woman with fair skin and light brown hair in two braids, wearing a black bra. She is performing oral sex on a man, sucking his erect, circumcised penis. Her lips are wrapped around the head of the penis, and her tongue is visible."
"BlowJobRK, Photograph of a woman with light skin and blonde hair in a bun, wearing gray headphones. She is topless with large breasts and a tattoo on her right shoulder. She is kneeling and performing oral sex on a man, with her lips wrapped around his penis."
"BlowJobRK, Photograph of a woman with pale skin and black hair with bangs, kneeling on a patterned rug. She is topless and wears a black choker with a star pendant. She is performing oral sex on a man, looking up at the camera. The camera angle is from above."
Training
- Base model: Krea 2
- Type: Concept LoRA
- Steps: 4000
- Rank: 32
- Optimizer: adamw, constant learning rate
- Engine: ai-toolkit
Generation settings
Trained on Krea 2, and works on Krea 2 Turbo for faster generation.
- Raw: ~30 steps, guidance 3.5
- Turbo: ~8 steps, guidance 1.0
The gallery images were generated with krea2_turbo_mxfp8.safetensors, at 8 steps, guidance 1, LoRA weight 1.



















