Z-IMAGE BASE Manual Install Guide V2

Manual Install Method from my last Youtube video: https://youtu.be/PAh870We7tI

Windows Installation
Step 1: Prerequisites

Make sure you have the following installed:

7-Zip → https://www.7-zip.org/

Git → https://git-scm.com/download/win

Step 2: Download and Extract ComfyUI

Download the ComfyUI NVIDIA portable build (v0.11.1):

https://github.com/comfyanonymous/ComfyUI/releases/download/v0.11.1/ComfyUI_windows_portable_nvidia.7z

Extract it with 7-Zip.

Your folder should look like:

ComfyUI_windows_portable/
├── ComfyUI/
│   ├── models/
│   └── custom_nodes/
├── python_embeded/
└── run_nvidia_gpu.bat

Step 3: Choose your Z-IMAGE quantization

You will download 2 UNets (Z-IMAGE + Z-IMAGE Turbo) in one of these versions:

Q5_K_S -> GPUs under 8 GB VRAM recommended

Q6_K -> GPUs 8 to 12 GB

Q8_0 -> Best quality, GPUs 12 to 16 GB and more

Remember your choice because the filenames depend on it.

Step 4: Install Required Custom Nodes

Open cmd inside:

ComfyUI_windows_portable\ComfyUI\custom_nodes

Run:

git clone https://github.com/ltdrdata/ComfyUI-Manager.git
git clone https://github.com/city96/ComfyUI-GGUF.git
git clone https://github.com/rgthree/rgthree-comfy.git
git clone https://github.com/yolain/ComfyUI-Easy-Use.git
git clone https://github.com/kijai/ComfyUI-KJNodes.git
git clone https://github.com/ssitu/ComfyUI_UltimateSDUpscale.git
git clone https://github.com/cubiq/ComfyUI_essentials.git
git clone https://github.com/wallish77/wlsh_nodes.git
git clone https://github.com/vrgamegirl19/comfyui-vrgamedevgirl.git
git clone https://github.com/ClownsharkBatwing/RES4LYF.git
git clone https://github.com/Jonseed/ComfyUI-Detail-Daemon
git clone https://github.com/Fannovel16/comfyui_controlnet_aux

Step 5: Install dependencies for each node

Still in:

ComfyUI_windows_portable\ComfyUI\custom_nodes

Run these (only for nodes that have a requirements.txt):

..\..\python_embeded\python.exe -m pip install -r ComfyUI-Manager\requirements.txt
..\..\python_embeded\python.exe -m pip install -r ComfyUI-GGUF\requirements.txt
..\..\python_embeded\python.exe -m pip install -r rgthree-comfy\requirements.txt
..\..\python_embeded\python.exe -m pip install -r ComfyUI-Easy-Use\requirements.txt
..\..\python_embeded\python.exe -m pip install -r ComfyUI-KJNodes\requirements.txt
..\..\python_embeded\python.exe -m pip install -r ComfyUI_UltimateSDUpscale\requirements.txt
..\..\python_embeded\python.exe -m pip install -r ComfyUI_essentials\requirements.txt
..\..\python_embeded\python.exe -m pip install -r wlsh_nodes\requirements.txt
..\..\python_embeded\python.exe -m pip install -r comfyui-vrgamedevgirl\requirements.txt
..\..\python_embeded\python.exe -m pip install -r RES4LYF\requirements.txt
..\..\python_embeded\python.exe -m pip install -r ComfyUI-Detail-Daemon\requirements.txt
..\..\python_embeded\python.exe -m pip install -r comfyui_controlnet_aux\requirements.txt


If one fails, keep going and install the rest first. Then retry the failed one.

Step 6: Download Z-IMAGE model files

All files are hosted here:

👉 https://huggingface.co/Aitrepreneur/FLX/tree/main

Place them under the correct folders in:

ComfyUI_windows_portable\ComfyUI\models\

Text Encoder

Put this file in:
models\text_encoders\

Qwen3-4B-UD-Q6_K_XL.gguf

VAE

Put this file in:
models\vae\

ae.safetensors

UNets (your chosen quantization)

Put these files in:
models\unet\

Pick ONE quantization (Q5_K_S or Q6_K or Q8_0) and download both of these matching files:

z_image_turbo-<MODEL_VERSION>.gguf

z_image-<MODEL_VERSION>.gguf

Examples:

If you chose Q6_K:

z_image_turbo-Q6_K.gguf

z_image-Q6_K.gguf



MODEL PATCHES
Z-Image-Fun-Controlnet-Union-2.1.safetensors

LORAS
Z-Image-Fun-Lora-Distill-8-Steps.safetensors



Upscalers

Put these files in:
models\upscale_models\

4x-ClearRealityV1.pth

RealESRGAN_x4plus_anime_6B.pth

Step 7: Run ComfyUI

Double-click:

ComfyUI_windows_portable\run_nvidia_gpu.bat










RunPod Installation
1. Create account at: https://bit.ly/runpodAi

2. Deploy a Pod

Pick a GPU that fits your quantization choice (Q8_0 is heavier)

Use a ComfyUI template (example: aitrepreneur/comfyui if that’s what you use in your videos)

Open the ComfyUI port (commonly 8188 on many templates, but use whatever your template exposes)

3. Go to your ComfyUI root folder

You must be inside a folder that contains:

models/

custom_nodes/

You can verify with:

ls

4. Create and activate a venv

Inside the ComfyUI root:

python3 -m venv venv
source venv/bin/activate
pip install -U pip setuptools wheel

5. Install Torch (recommended baseline that matches the script)

Your script pins Torch to 2.4.0 with cu121 by default.

pip install --index-url https://download.pytorch.org/whl/cu121 --extra-index-url https://pypi.org/simple \
  torch==2.4.0+cu121 torchvision==0.19.0+cu121 torchaudio==2.4.0+cu121


Then install the common stability pins:

pip uninstall -y opencv-python || true
pip install opencv-python-headless==4.12.0.88
pip install "pillow>=11.0.0"
pip install gguf piexif || true
pip install librosa || true

6. Clone the custom nodes

From the ComfyUI root:

cd custom_nodes

git clone https://github.com/ltdrdata/ComfyUI-Manager.git
git clone https://github.com/city96/ComfyUI-GGUF.git
git clone https://github.com/rgthree/rgthree-comfy.git
git clone https://github.com/yolain/ComfyUI-Easy-Use.git
git clone https://github.com/kijai/ComfyUI-KJNodes.git
git clone https://github.com/ssitu/ComfyUI_UltimateSDUpscale.git
git clone https://github.com/cubiq/ComfyUI_essentials.git
git clone https://github.com/wallish77/wlsh_nodes.git
git clone https://github.com/vrgamegirl19/comfyui-vrgamedevgirl.git
git clone https://github.com/ClownsharkBatwing/RES4LYF.git
git clone https://github.com/Jonseed/ComfyUI-Detail-Daemon
git clone https://github.com/Fannovel16/comfyui_controlnet_aux
cd ..

7. Install node requirements

From the ComfyUI root:

pip install -r custom_nodes/ComfyUI-Manager/requirements.txt || true
pip install -r custom_nodes/ComfyUI-GGUF/requirements.txt || true
pip install -r custom_nodes/rgthree-comfy/requirements.txt || true
pip install -r custom_nodes/ComfyUI-Easy-Use/requirements.txt || true
pip install -r custom_nodes/ComfyUI-KJNodes/requirements.txt || true
pip install -r custom_nodes/ComfyUI_UltimateSDUpscale/requirements.txt || true
pip install -r custom_nodes/ComfyUI_essentials/requirements.txt || true
pip install -r custom_nodes/wlsh_nodes/requirements.txt || true
pip install -r custom_nodes/comfyui-vrgamedevgirl/requirements.txt || true
pip install -r custom_nodes/RES4LYF/requirements.txt || true
pip install -r custom_nodes/ComfyUI-Detail-Daemon/requirements.txt || true
pip install -r custom_nodes/comfyui_controlnet_aux/requirements.txt || true


(Using || true keeps the process going if one node is annoying. You can retry the failed one after.)

8. Download Z-IMAGE model files (with curl)

All files come from:

https://huggingface.co/Aitrepreneur/FLX/resolve/main

Pick your quantization:

MODEL_VERSION=Q5_K_S or Q6_K or Q8_0

Example (Q8_0):

MODEL_VERSION=Q8_0
HF_BASE="https://huggingface.co/Aitrepreneur/FLX/resolve/main"

mkdir -p models/text_encoders models/vae models/unet models/upscale_models

curl -L -o "models/text_encoders/Qwen3-4B-UD-Q6_K_XL.gguf" "$HF_BASE/Qwen3-4B-UD-Q6_K_XL.gguf?download=true"
curl -L -o "models/vae/ae.safetensors" "$HF_BASE/ae.safetensors?download=true"

curl -L -o "models/unet/z_image_turbo-${MODEL_VERSION}.gguf" "$HF_BASE/z_image_turbo-${MODEL_VERSION}.gguf?download=true"
curl -L -o "models/unet/z_image-${MODEL_VERSION}.gguf" "$HF_BASE/z_image-${MODEL_VERSION}.gguf?download=true"

curl -L -o "models/upscale_models/4x-ClearRealityV1.pth" "$HF_BASE/4x-ClearRealityV1.pth?download=true"
curl -L -o "models/upscale_models/RealESRGAN_x4plus_anime_6B.pth" "$HF_BASE/RealESRGAN_x4plus_anime_6B.pth?download=true"

curl -L -o "models/model_patches/Z-Image-Fun-Controlnet-Union-2.1.safetensors" "$HF_BASE/Z-Image-Fun-Controlnet-Union-2.1.safetensors?download=true"
curl -L -o "models/loras/Z-Image-Fun-Lora-Distill-8-Steps.safetensors" "$HF_BASE/Z-Image-Fun-Lora-Distill-8-Steps.safetensors?download=true"




9. Run ComfyUI and watch logs

Common patterns (depends on template):

tail -f /workspace/logs/comfyui.log


If your template uses a start script, use that. Otherwise, you typically run ComfyUI from its main folder while venv is active.