Photoreal character for Z-Image Turbo · LoRA by SeedraAI
A consistent fictional character for Z-Image Turbo, made for realistic phone-photo content. The trigger alone is not enough: always put the full appearance block from the prompt template at the start, otherwise the base model takes over and you get a different person.
Works well on: portraits and medium shots, everyday phone-photo scenes.
AIZK1s
Prompt template:
AIZK1s, young East Asian woman, very long straight glossy black hair with center part falling past the chest, oval face, high cheekbones, almond-shaped dark brown eyes with slightly hooded lids, thin straight nose, full nude-toned lips, thin soft eyebrows, smooth warm beige skin, slim build with narrow shoulders, {your scene}
| LoRA weight | 1.0 |
| Size | portrait, about 1 to 2 MP |
| Steps / CFG | Turbo defaults |
| Sampler | Turbo defaults |
AIZK1sZIMAGE.safetensors into ComfyUI/models/loras/.Every preview here was made locally in ComfyUI with this LoRA. The full SeedraAI workflows behind them (realism stack, camera noise, grain and color post in one graph) are part of the SeedraAI membership: seedraai.com.
Tongyi-MAI/Z-Image-Turbo