cleavegagged

Produces images of women gagged with cloth between their lips. This model aims to produce slightly thicker, often knotted gags compared to some models which produce very thin gags. The dataset focused on very pretty women wearing gags that were flattering to their lips.

Works best with simple prompts in a face closeup or portrait shot format.

For longer or more complicated prompts, inpainting might be necessary to fix the gag/face once the desired pose/background/clothing comes out right.

Tends to respond best to high LoRA strength, 1.0-1.2 recommended. Lowering strength improves flexibility but can remove or alter the gag, which can hopefully be fixed in img2img or inpaint.

Color of cloth can be specified with tag: gagged with (color of choice) cloth. For scarf/bandana style gags, try "gagged with color of choice patterned cloth."