Z-image turbo workflow | CivitAI metadata, upscaling, controlnet, FaceDetailer (+indexing elements), inpainting ComfyUI beginner friendly — v2

Use this ComfyUI workflow for Z-Image Turbo generation with ControlNet, upscaling, FaceDetailer and inpainting; its graph includes text-encoding and image-loading inputs, with image-preview and image-saving outputs.

The listed model-file roles are diffusion model, text encoder, VAE and ControlNet; no saved-image format is named.

Version v2 lists ZImageTurbo as its base model.

Listed model files: bbox/face_yolov8m.pt; Z-Image-Turbo-Fun-Controlnet-Union-2.1-lite-2602-8steps.safetensors; z-image-turbo-q8_0.gguf.

Other listed model files: ema_vae_fp16.safetensors; pig_flux_vae_fp32-f16.gguf; Qwen_Qwen3-4B-Instruct-2507-Q6_K_L.gguf; seedvr2_ema_7b-Q4_K_M.gguf.

Listed node packs: ComfyLiterals; ComfyUI_Comfyroll_CustomNodes; comfyui_controlnet_aux; comfyui_smznodes; comfyui-custom-scripts; comfyui-image-saver; comfyui-impact-pack; comfyui-inpaint-cropandstitch; rgthree-comfy.

The graph includes selectors for aspect ratio, sampler and scheduler.

The workflow lists comfyui_controlnet_aux at version 1.0.7; verify that package and version are available in your ComfyUI registry before setup.

Suggestion · not verified

No license is listed in the source details; check the Civitai usage terms before reusing the workflow.

Suggestion