Z-Image Turbo + Controlnet (with LoRA fix) + 4k Upscaling + Detail Daemon — v1.0

This workflow creates Z-Image Turbo images using ControlNet conditions and LoRAs, with smoothness and Detail Daemon controls and optional 4K upscaling through SeedVR2.

Its dual-sampler setup applies LoRAs at low strength in early steps, then at full strength in the final steps.

Available ControlNet conditions include Canny, Depth, Pose, MLSD, HED, Scribble, and Gray; SAM3 can isolate a character or object as a pose reference.

You can use up to eight LoRA slots and adjust their strengths.

For a more basic GGUF-compatible workflow, the description points to “ZIT standard workflow.”

Suggestion

The workflow lists these node packs: cg-use-everywhere, comfyui_controlnet_aux, comfyui_essentials, comfyui_fill-nodes, comfyui_layerstyle, comfyui_lg_samplingutils, comfyui-custom-scripts, comfyui-detail-daemon, comfyui-easy-use, comfyui-impact-pack, comfyui-kjnodes, comfyui. .

The workflow’s file list names ae.safetensors, DigitalAnime.safetensors, dw-ll_ucoco_384_bs5.torchscript.pt, ema_vae_fp16.safetensors, qwen_3_4b.safetensors, seedvr2_ema_7b_sharp_fp16.safetensors, and TP2KKFHC9NPFNVSK522R051N20.safetensors.

It also names woman037-zimage.safetensors, woman549-zit.safetensors, woman992-ZIT.safetensors, yolox_l.onnx, z_image_bf16.safetensors, and Z-Image-Turbo-Fun-Controlnet-Union.safetensors.

Yellow nodes mark editable inputs and settings; red nodes contain instructions and notes.

Smoothness and Detail Daemon provide controls for detail in the final image.

The listed details provide no download locations for DigitalAnime.safetensors, dw-ll_ucoco_384_bs5.torchscript.pt, ema_vae_fp16.safetensors, seedvr2_ema_7b_sharp_fp16.safetensors, TP2KKFHC9NPFNVSK522R051N20.safetensors, woman549-zit.safetensors, and woman992-ZIT.safetensors; 확인rì

Suggestion

Before setup, verify which packages your ComfyUI installation should use for comfyui_controlnet_aux and seedvr2_videoupscaler.

Suggestion