All-in-One. Z-Image, Ernie, Klein, t2i, i2i, controlNet, inpaint, outpaint, LLM, WAN, Qwen image, SDXL, Chroma — v8.9

This ComfyUI workflow is for creating and editing images from text prompts and one or more input images; it supports txt2img, img2img, inpainting, outpainting, face swap, image description, prompt enhancement, and SeedVR2 upscaling, producing generated, edited, inpainted, outp​​ת

Listed model options include 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; ControlNet is listed only for Edit models in the img2img group, where the prompt can request an input image

Use a text prompt for txt2img, or load one or more images for img2img and Edit/Kontext; Edit/Kontext supports loading a single image or multiple images.

The listed operations produce generated or edited images, inpainted and outpainted images, face-swapped images, upscaled images, image descriptions, and enhanced prompts.

Before installing custom nodes through the manager, install ComfyUI-ppm from https://github.com/BigStationW/ComfyUI-ppm; the listing associates it with negative prompts for Flux1, ZIT, and Anima.

Vorschlag · nicht geprüft

If the manager cannot find the LLM or LoRA nodes, install ComfyUI-Lora-Manager from https://github.com/willmiao/ComfyUI-Lora-Manager and use https://github.com/KingManiya/ComfyUI-LLM-text-processor for the LLM node; automatic setup for the latter is described only for Windows x64

Vorschlag · nicht geprüft

Check whether these listed node packs are available before use: basic_data_handling; cg-use-everywhere; chibi; comfyui_controlnet_aux; rgthree-comfy; comfyui_facedetectionnode; comfyui-inpaint-cropandstitch; comfyui-crystools; comfyui-custom-scripts; comfyui-easy-use; comfyui-gg

Vorschlag · nicht geprüft

Check for these listed model files: umt5-xxl-encoder-Q5_K_S.gguf; qwen_3_06b_base.safetensors; qwen_image_vae.safetensors; Magic-Wan-Image-v2-Q4_1.gguf; qwen_2.5_vl_7b_fp8_scaled.safetensors; segm/person_yolov8m-seg.pt; Krea2_Turbo_fp8mixed.safetensors; ernie-image-Q6_K.gguf; mm

Vorschlag · nicht geprüft

Also check these listed files and components: z_image_turbo-Q8_0.gguf; ema_vae_fp16.safetensors; qwen3vl_4b_fp8_scaled.safetensors; t5-v1_1-xxl-encoder-bf16.safetensors; flux1-dev-Q5_1.gguf; Chroma1-HD-fp8_scaled_rev2.safetensors; qwen_3_4b.safetensors; ministral-3-3b.safetensors

Vorschlag · nicht geprüft

Put unchanging instructions in Fixed and editable instructions in User Prompt; when LLM processing is enabled, Fixed stays unchanged and only User Prompt can be improved.

Vorschlag · nicht geprüft

Starting from v8.0, use the T2I switches to turn prompt enhancement on the fly; prompts from the LLM group still need to be copied manually into the I2I group, while Image 1 in the I2I block can be described directly.

Vorschlag · nicht geprüft

Sage Attention is listed as usable through ComfyUI launch arguments; the listing points to https://github.com/wildminder/AI-windows-whl?tab=readme-ov-file and says no compiler is needed for that installation.

Vorschlag · nicht geprüft

Face swap is described as “not that perfect,” so review its result; before downloading, verify the page title is “All-in-One. Z-Image, Ernie, Klein, t2i, i2i, controlNet, inpaint, outpaint, LLM, WAN, Qwen image, SDXL, Chroma”, the publisher is TikFesku, and the URL is https://c

Vorschlag · nicht geprüft

The listed parts are not confirmed as connected, so verify the node and model setup before use.

Vorschlag · nicht geprüft