Have a ComfyUI workflow? Get a readable Workflow Doc in your browser.Make a workflow readable.Try it now ↗
This is a workflow that I cobbled together for Z-Image. It uses Z-Image base for composition, Z-Image Turbo for detailing and then a Z-Image Turbo skin detailing pass. I don't claim this workflow is the most...
Runtime profile
Source description
This is a workflow that I cobbled together for Z-Image. It uses Z-Image base for composition, Z-Image Turbo for detailing and then a Z-Image Turbo skin detailing pass. I don't claim this workflow is the most efficient or well thought out but it has given me good results, YMMV.
You can use LORAs in the generation stage and detailing stages. I set the ModelSamplingAuraFlow to 7-9 for composition. In the detailing stage use denoise as required but I usually have it between .28-.34. For the skin detailing pass I leave it at .32 and it gives me good results.
If you don't like LORA Manager you can swap out those nodes. Also, I use subgraphs to help organize the workflow but feel free to unpack those to your liking!
AI-generated commentary
AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.
Berts Eyebrow's Workflow (Z-image) is for creating Z-Image images: it uses Z-Image base for composition, adds Z-Image Turbo detailing and a separate skin-detailing pass, accepts optional LORAs during generation or detailing, and produces the resulting detailed image.
The flow moves from Z-Image composition to Z-Image Turbo detailing, followed by a Z-Image Turbo skin-detailing pass.
Subgraphs organize the workflow's stages.
Optional LORAs can be used during both generation and detailing stages.
The output is a Z-Image result after composition, detailing, and a separate skin-detailing pass.
This workflow is titled Berts Eyebrow's Workflow (Z-image), is published by Berts Eyebrow, and uses ZImageBase as its base model.
Listed model files are ae.safetensors, z_image_bf16.safetensors, z_image_turbo_bf16.safetensors, qwen_3_4b.safetensors, and RealESRGAN_x4plus.safetensors.
Two additional model files are listed: bigLove_zt3.safetensors and 4xFaceUpSharpDAT.safetensors.
Listed node packs are comfy-image-saver, comfyui-easy-use, comfyui-kjnodes, comfyui-lora-manager, and controlaltai-nodes.
For composition, the author sets ModelSamplingAuraFlow to 7–9.
Suggestion · not verified
For detailing, the author usually uses denoise .28–.34 and leaves the skin-detailing pass at .32.
Suggestion · not verified
LORA Manager nodes can be replaced, and subgraphs can be unpacked if preferred.
Suggestion · not verified
Before using it, verify that bigLove_zt3.safetensors and 4xFaceUpSharpDAT.safetensors are available; neither file has an identified provider.
Suggestion · not verified
No usage license is listed, so check the source page for usage terms before using the workflow.
Suggestion · not verified
Does this need to be edited?
Sign in to send an edit request.
Sources
1 sourceSource excerpts
2 excerptsSource context: 123 downloads · Type Workflows · Base model ZImageBase
Estimated VRAM requirement
Estimate unavailable
Model file sizes are not complete enough to calculate this requirement yet.
Requirements
12 requirements4xFaceUpSharpDAT.safetensors
Not resolvedUpscaler · Unknown
VAE · Hugging Face · Comfy-Org/z_image_turbo
bigLove_zt3.safetensors
Not resolvedCheckpoint · Unknown
Text encoder · Hugging Face · Comfy-Org/z_image_turbo
Upscaler · Hugging Face · Comfy-Org/Real-ESRGAN_repackaged · Real ESRGAN Repackaged
Checkpoint · Hugging Face · Comfy-Org/z_image
z_image_turbo_bf16.safetensors
File unverifiedCheckpoint · Hugging Face · Comfy-Org/z_image_turbo
Node pack · Registry
Node pack · Registry
Source context: 121 downloads · Type Workflows · Base model ZImageBase
Node pack · Registry
Node pack · Registry
Node pack · Registry
