POV Handjob - Z Image Turbo (Photo & Anime)

Details

Download Files

Model description

A LoRA trained to generate first-person perspective handjob compositions from the man's point of view. This model was trained on a curated dataset of over 58 images with light captioning. Generates a decent looking penis a lot of the time compared to base Z-Image. Works well for photo style gens and surprisingly very well for Anime considering there was none in the training set.

Main Trigger Word:

povhj (essential for activating the primary positioning logic and camera perspective). While it will produce results without the trigger, using povhj is strongly recommended for consistent composition and framing.

Core Conceptual Keywords:

  • Positioning (THESE ARE PRETTY KEY): a woman giving a handjob, one hand on penis, two hands on penis, kneeling between his legs, kneeling on the ground, woman standing beside the bed. The trigger structure clearly defines the woman's position relative to the camera and subject (kneeling between his legs,standing beside the bed), allowing for reliable control over the core action.

  • Framing: Man`s spread legs and penis at the bottom of the frame, Just a penis visible at the bottom of the frame, man is standing, just his penis and legs visible at bottom of frame.

  • Action/Detail: licking tip of penis, breast pressed against penis, looking at camera, looking at penis.

  • Body/Face: curvy body, soft body, skinny body, petite body | huge breasts, large breasts, medium breasts, small breasts, | good about accepting most age or appearance descriptions.

  • Clothing: Can handle [clothing item] pushed up/pulled down decently.

What it's capable of:

  • Consistent POV Framing: Excels at maintaining the man's perspective, reliably positioning the penis at the bottom of the frame and the woman in the foreground. The specified framing keywords work with high consistency.

  • Archetype & Setting Variety: Can generate a wide range of female archetypes and integrate them into diverse settings/environments

  • Body Diversity: Trained on a comprehensive range of body types and breast sizes. Responds very well to descriptors for age, body shape, and unique features.

  • Some additional "sub-trainings": Try out: "She is sticking her tongue out licking the tip of the penis." & "leaning over him and touching his penis to her breast" & "She has cum on her face and breasts" (This one works inconsistently)

Known Quirks:

  • The lora performs best when the prompt structure is followed closely. Deviating too far from the established pattern can lead to inconsistent composition.

  • The exact strength may need to be adjusted (typically 0.7-1.0) to balance the pose with other stylistic elements in the prompt. But I find 1.0 works fine unless I really want try to break the mold

  • The man's and woman's legs can get mixed up but generally can be fixed by a re-gen or a slight prompt change.

Images made by this model

No Images Found.