Installation Guide

ComfyUI supports multiple installation methods. Choose the one that best fits your needs and technical level.

System Requirements

Minimum Requirements

• OS: Windows 10/11, macOS 10.15+, Linux
• RAM: 8GB
• GPU: NVIDIA GPU with 4GB+ VRAM (recommended)
• Storage: 20GB+ free space

Recommended Configuration

• RAM: 16GB+
• GPU: NVIDIA GPU with 8GB+ VRAM (RTX 3070 or higher)
• Storage: SSD for faster loading

Windows Installation

Method 1: Portable Version (Recommended for Beginners)

1. Visit ComfyUI GitHub Releases
2. Download the latest Windows portable archive (*_windows_amd64_cu121*.7z)
3. Extract to your desired location
4. Run run_nvidia_gpu.bat (NVIDIA GPU) or run_cpu.bat (CPU only)

Method 2: Git Clone

# Clone repository
git clone https://github.com/comfyanonymous/ComfyUI.git

# Enter directory
cd ComfyUI

# Install dependencies (recommended to create a virtual environment first)
pip install -r requirements.txt

Method 3: Using Manager

ComfyUI Manager provides a graphical interface for installation and updates:

1. Install ComfyUI (any method)
2. Download Manager:

   cd custom_nodes
   git clone https://github.com/ltdrdata/ComfyUI-Manager.git

3. Restart ComfyUI, you'll see the Manager entry in the interface

macOS Installation

Apple Silicon (M1/M2/M3)

# Clone repository
git clone https://github.com/comfyanonymous/ComfyUI.git
cd ComfyUI

# Create virtual environment (recommended)
python3 -m venv venv
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

Run: ./run_mps.sh or python3 main.py --force-fp16

Intel Mac

For Intel Mac, consider using RyzenAI or CPU-only mode.

Linux Installation

Basic Installation

# Clone repository
git clone https://github.com/comfyanonymous/ComfyUI.git
cd ComfyUI

# Install system dependencies (Ubuntu/Debian)
sudo apt update
sudo apt install python3 python3-pip python3-venv

# Create virtual environment and install
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

NVIDIA GPU Configuration

Ensure CUDA drivers are installed:

# Check driver
nvidia-smi

# Install CUDA (example: CUDA 12.1)
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.0-1_all.deb
sudo dpkg -i cuda-keyring_1.0-1_all.deb
sudo apt update
sudo apt install cuda-12-1

Verify Installation

After installation, start ComfyUI with:

• Windows: Double-click run_nvidia_gpu.bat
• macOS/Linux: Run ./run_nvidia_gpu.sh or ./run_cpu.sh

Visit http://127.0.0.1:8188 . If you see the ComfyUI interface, the installation is successful.

Initial Configuration

Download Models

ComfyUI doesn't include models by default. You need to download separately:

1. Base Models: Download from Civitai or Hugging Face
2. Placement: Usually in models/checkpoints/ directory

Model Directory Structure

ComfyUI/
├── models/
│   ├── checkpoints/      # Base models (SD1.5, SDXL)
│   ├── vae/              # VAE models
│   ├── loras/             # LoRA models
│   ├── controlnet/        # ControlNet models
│   ├── embeddings/        # Textual Inversion embeddings
│   └── upscale_models/    # Super-resolution models

Troubleshooting

Q: "CUDA out of memory" error on startup

A: Try reducing batch size or use smaller models. You can also enable "Low VRAM mode" in settings.

Q: Missing Visual Studio Build Tools

A: Windows users need to install Visual Studio Build Tools, select "C++ build tools".

Q: pip install fails

A: Ensure you're using Python 3.10-3.11, avoid Python 3.12.

Next Steps

After installation, continue reading:

• Interface Guide - Learn interface layout and basic operations
• Node Reference - Master node workflows