Manual Comfyui Kontext Models Install

Manual Install Method from my last Youtube Video: https://youtu.be/Lsz3tlbBaqM

Download and extract the ComfyUI portable version from:
https://github.com/comfyanonymous/ComfyUI

After extraction, your folder should look like this:

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

Install Required Custom Nodes
Navigate to:
ComfyUI_windows_portable/ComfyUI/custom_nodes

Then open terminal or CMD in this folder and run:

git clone https://github.com/city96/ComfyUI-GGUF
git clone https://github.com/rgthree/rgthree-comfy
git clone https://github.com/kijai/ComfyUI-KJNodes
git clone https://github.com/crystian/ComfyUI-Crystools
git clone https://github.com/cubiq/ComfyUI_essentials
git clone https://github.com/wallish77/wlsh_nodes
git clone https://github.com/ChenDarYen/ComfyUI-NAG

IMPORTANT: For any folder that has a requirements.txt, install dependencies using:
..\..\python_embeded\python.exe -m pip install -r PATH\TO\requirements.txt

Download Kontext Model Files
All model files are hosted here:
https://huggingface.co/Aitrepreneur/FLX/tree/main

for the models:

in the text_encoders folder:
t5xxl_fp8_e4m3fn_scaled.safetensors

in the clip folder:
clip_l.safetensors

in the vae folder:
ae.safetensors

IF THE USER SELECTED YES TO "FP8 models" then in the diffusion_models folder:
flux1-kontext-dev-fp8-e4m3fn.safetensors

if you want the nunchaku models:
svdq-int4_r32-flux.1-kontext-dev.safetensors   ---> older than 5000 series
svdq-fp4_r32-flux.1-kontext-dev.safetensors    --->5000 series


depending on the model that the user selected at the beginning, in the unet folder:
flux1-kontext-dev-Q4_K_S.gguf
flux1-kontext-dev-Q5_K_S.gguf
flux1-kontext-dev-Q8_0.gguf

in the loras folder:
schnell_v1.0.safetensors


in the upscale_models folder
4x-ClearRealityV1.pth
RealESRGAN_x4plus_anime_6B.pth




TO INSTALL NUNCHAKU:


Navigate to your ComfyUI_windows_portable\python_embeded directory

Uninstall problematic previous dependencies: This cleans up any prior failed attempts or conflicting versions.
python.exe -m pip uninstall nunchaku insightface facexlib filterpy diffusers accelerate onnxruntime -y


Install the specific Nunchaku development wheel: This is crucial as it's a pre-built package that bypasses common compilation issues and is compatible with PyTorch 2.7 and Python 3.12.

python.exe -m pip install https://github.com/mit-han-lab/nunchaku/releases/download/v0.3.1dev20250609/nunchaku-0.3.1.dev20250609+torch2.7-cp312-cp312-win_amd64.whl


Install facexlib: After installing the Nunchaku wheel, the facexlib dependency for some optional nodes (like PuLID) might still be missing. Install it directly.

python.exe -m pip install facexlib


Install insightface: insightface is another crucial dependency for Nunchaku's facial features. It might not be fully pulled in by the previous steps.

python.exe -m pip install cython


then download this file: https://huggingface.co/Aitrepreneur/FLX/resolve/main/insightface-0.7.3-cp312-cp312-win_amd64.whl?download=true

python.exe -m pip install insightface-0.7.3-cp312-cp312-win_amd64.whl



Install onnxruntime: insightface relies on onnxruntime to run ONNX models. Ensure it's installed.

python.exe -m pip install onnxruntime

python.exe -m pip install onnxruntime-gpu



Then once it's done:

install the custom node inside the ComfyUI_windows_portable\ComfyUI\custom_nodes folder:
https://github.com/mit-han-lab/ComfyUI-nunchaku

then install the requirements:
..\..\python_embeded\python.exe -m pip install -r ComfyUI-nunchaku\requirements.txt



Run ComfyUI




## RunPod Installation

1. Create account at: https://bit.ly/runpodAi
2. Choose a 24GB VRAM GPU with the aitrepreneur/comfyui template
3. Access port 8888
4. Follow the same folder structure as above
5. Use `curl` to download files:

   curl -L -o "filename" "URL"

6. Monitor logs with:

   tail -f workspace/logs/comfyui.log