no custom nodes - official ComfyUI template with 3/4 manually added LoRAs - easy install THIS workflow > {OR try the advanced workflow Lora Loader version [requires custom nodes]: https://civitai.com/articles/31802/lora-loader-workflow-z-image-turbo } (v2 Requires ComfyUI v0.27.0 for INT8 Support)
Models:
Z Image Turbo INT8 [diffusion_model] { https://huggingface.co/Comfy-Org/z_image_turbo/blob/main/split_files/diffusion_models/z_image_turbo_int8_convrot.safetensors }
Qwen 3 4B INT8 [clip/text_encoder] { https://huggingface.co/Winnougan/Z-Image-Base-Turbo-INT8-Convrot/blob/main/qwen_3_4b_int8_convrot.safetensors }
ComfyUI default VAE [decoder] { https://huggingface.co/Comfy-Org/z_image_turbo/blob/main/split_files/vae/ae.safetensors }
HI8 [LoRA@0.04] { https://civitai.com/models/1134895/2000s-analog-core?modelVersionId=2456725 }
Olympus [LoRA@0.20] { https://civitai.com/models/2212121/olympus-ultrareal?modelVersionId=2490572 }
Z-Detail-Slider [LoRA@0.69-1.00] { https://civitai.com/models/2234266/zit-detail-slider?modelVersionId=2515203 }
NSFW_master_ZIT [LoRA@0.07] { https://civitai.com/models/667086/nsfw-master?modelVersionId=2904324 } (off)
NSFW LoRA used at extremely low weight (strength_model: 0.07) for additional detail, adjust between 1.00 and 3.00 for actual NSFW output, with 2.00 being a balanced middle-ground of prompt adherence and overall image quality.
v1:
Models:
Ultra Flux 4k VAE { https://huggingface.co/Owen777/UltraFlux-v1/blob/main/vae/diffusion_pytorch_model.safetensors } OR [ae.safetensors] provided in ComfyUI's default template workflow
Jib Mix ZIT [fp16] { https://civitai.com/api/download/models/2511897?fileId=2399888 }
qwen_3_4b.safetensors can be downloaded in ComfyUI desktop
HI8 [@0.04] { https://civitai.com/models/1134895/2000s-analog-core?modelVersionId=2456725 }
Olympus [@0.20] { https://civitai.com/models/2212121/olympus-ultrareal?modelVersionId=2490572 }
Z-Detail-Slider [@0.69-1.00] { https://civitai.com/models/2234266/zit-detail-slider?modelVersionId=2515203 }
NSFW_master_ZIT [@0.07] { https://civitai.com/models/667086/nsfw-master?modelVersionId=2904324 } (off)
VAE comparison:
if you dislike the added detail provided by UltraFlux4k vae, feel free to use ae.safetensors for a more smoothed effect across the entire output image — image below uses Ultra Flux
the effect is most noticeable around the edges of the DJ's face and the 4 lines in the middle of the DJ Deck. The top first image does not use Ultra Flux and instead uses ae.safetensors, so it doesn't have the small white line along the right side of the image, and instead has a bit of a blurry photoshopped look filter across the image compared to the 2nd's crisp detail.
[very low resolution images for example]
