All-in-One. Z-Image, Ernie, Klein, t2i, i2i, controlNet, inpaint, outpaint, LLM, WAN, Qwen image, SDXL, Chroma — v6.1

This ComfyUI workflow is for creating and editing images from text prompts and one or more input images; it can produce text-to-image, image-to-image, inpainted, and outpainted results, plus face swaps, SeedVR2 upscaling, Edit-model ControlNet use, and LLM-assisted image captions

It supports Z-Image, Krea-2, Ernie image, MS Lens, Flux, Flux 2, Flux 2 Klein, Anima, Qwen Image, Qwen Image Edit, SDXL/Pony, Chroma, and Lumina-Image 2.0.

Outpainting is used in Edit/Kontext mode, while ControlNet is limited to Edit models in the image-to-image group.

Enter a text prompt for text-to-image, or load one image or multiple images for image-to-image and Edit/Kontext work.

Each prompt group has a Fixed section that stays at the beginning; when enabled, the LLM can improve the User Prompt, and image-to-image prompts must be copied manually from the LLM group.

Install ComfyUI-ppm from Git before installing custom nodes through the manager; if the manager cannot find the LLM or LoRA nodes, install ComfyUI-Lora-Manager from Git.

Suggestion · non vérifié

Automatic ComfyUI-LLM-text-processor setup supports Windows x64 with CUDA 13; other platforms require manual installation of llama.cpp release binaries.

Named node packs are comfyui-gguf, comfyui-lora-manager, rgthree-comfy, cg-use-everywhere, comfyui-easy-use, RES4LYF, comfyui-kjnodes, crt-nodes, comfyui-crystools, and comfyui-custom-scripts.

Named model files are qwen_3_4b.safetensors, Qwen3-4B-Q8_0.gguf, z_image_turbo_bf16.safetensors, and z_image_turbo-Q8_0.gguf.

For LLM use, the workflow names Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q8_0.gguf and mmproj-Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-BF16.gguf for 8 Gb VRAM; for 12+ Gb VRAM, it names Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_M.gguf with mmproj-Qwen3.6-35B-A3B-Unc全s

For LLM features, place model files in ComfyUI/models/LLM and system-prompt files in ComfyUI/models/LLM/prompts, including LLM-system.txt and LLM-system-SDXL.txt.

Suggestion · non vérifié

Sage Attention can be enabled in the ComfyUI launch arguments; for Anima, ZIT, and Flux1, the Fixed prompt can use (glasses, tie:-1) as an example for excluding glasses and ties.

Suggestion · non vérifié

Face swapping may be imperfect, so review those results before use.

Suggestion · non vérifié

Before setup, verify the exact names and presence of every node pack and model file listed for this workflow in your ComfyUI installation.

Suggestion · non vérifié