This is ZIT PONGO, my first LoRA trained on Z-Image. The style recreates a childish playdough look with soft shapes, visible fingerprints, and simple plastiline textures.
It performs best at weight 1.0.
I trained and mixed four separate LoRAs using Ostris-Ai Toolkit, then merged them with a custom merger built specifically for Z-Image.
It doesn’t require a trigger word, but the prompt format included with the gallery images delivers the most consistent results.