ComfyUI Installation Guide

In this section, we’ll go through several ways to install ComfyUI. All methods ultimately give you the same result, so you can simply choose the one that best matches your operating system and experience level.

Method 1: Download from the Official Website (Desktop Version)

Steps

1. Visit the official ComfyUI website: https://comfy.org/
2. Click the Download button and select the version for your operating system (Windows, macOS, or Linux).
3. Open the downloaded file and start the installation process.

During installation, the program will automatically download the required files and set up ComfyUI on your computer.

Important Notes

• Windows users: ComfyUI works best with NVIDIA GPUs because most AI models rely on CUDA acceleration.
• Automatic dependency installation: The installer will attempt to download and configure required dependencies automatically. However, in some environments you may still encounter missing packages or version conflicts.

Method 2: Download the Portable Version

Another easy way to run ComfyUI is to use the portable version. This version does not require a traditional installation.

Steps

1. Download the portable package from a trusted source.
2. Extract the archive to any folder on your computer.
3. Open the extracted folder.
4. Run the startup script (for example run_nvidia_gpu.bat on Windows).

The portable version already includes most required components, so it is often the quickest way to get started.

Method 3: Clone from the GitHub Repository (Recommended for Developers)

If you prefer more control over your environment, you can install ComfyUI directly from the source repository.

Steps

1. Make sure Git and Python are installed on your system. Recommended Python versions: 3.10 – 3.11.
2. Open a Terminal (Linux/macOS) or Command Prompt / PowerShell (Windows).
3. Run the following commands:

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

# Enter the project directory
cd ComfyUI

# Install required dependencies
pip install -r requirements.txt

This method works on Windows, macOS, and Linux.

Method 4: Windows Launcher (Beginner-Friendly)

Some third-party launchers provide a graphical interface that simplifies installation and environment configuration. These tools can automatically install Python, CUDA, and other dependencies.

Steps

1. Visit the download page: https://pan.quark.cn/s/64b808baa960
2. Download one of the packages (for example ComfyUI-aki-v2.7z or ComfyUI-aki-v3.7z).
3. Extract the archive to a folder on your computer.
4. Run the launcher executable file.
5. The launcher will detect your system environment and help install any missing components automatically.

This method is particularly convenient for users who prefer not to use the command line.

Method 5: Use a Cloud Platform

If your computer does not have a powerful GPU, you can also run ComfyUI on cloud platforms.

Some services provide ready-to-use environments, including:

• runninghub.com
• cloud.comfy.org
• liblib.ai(for chinese user)

Cloud platforms usually provide GPU resources, but they often come with usage limits or credit systems, so long-term use may require payment.

Method 6: use cnb platform

Running ComfyUI on CNB by Forking the Repository

This guide explains how to run ComfyUI directly on CNB by forking the repository and launching it in the cloud development environment. This approach requires no local installation and allows you to run ComfyUI entirely in a cloud workspace.

1. What is CNB

Cloud Native Buildpacks (CNB) is a cloud development environment platform that allows developers to:

• Fork Git repositories
• Automatically build environments
• Run applications directly in the browser
• Deploy applications without manual environment setup

CNB workspaces usually provide:

• Linux development environment
• Preinstalled development tools
• Public access URLs
• Persistent storage

This makes it convenient for running AI tools such as ComfyUI.

2. Prerequisites

Before starting, prepare the following:

1. CNB Account

Create an account on the CNB platform.

2. GitHub Account

You will need GitHub to fork the repository.

3. Basic Knowledge

Basic familiarity with:

• Linux terminal
• Git
• Python

3. Fork the ComfyUI Repository

Go to the official repository:

https://github.com/comfyanonymous/ComfyUI

Click Fork.

This creates your own copy of the repository in your GitHub account.

Example:

https://github.com/YOURNAME/ComfyUI

4. Import the Forked Repository into CNB

Open the CNB dashboard.

Steps:

1. Click New Workspace
2. Choose Import from Git Repository
3. Paste your forked repository URL

Example:

https://github.com/YOURNAME/ComfyUI

4. Select environment configuration.

Recommended environment:

OS: Ubuntu
Python: 3.10
CPU: 4 cores
RAM: 16GB

Then click Create Workspace.

CNB will automatically:

Clone repository
Create development environment
Start workspace

5. Open the CNB Terminal

Once the workspace is created, open the terminal inside CNB.

Navigate to the project directory:

cd ComfyUI

Verify files:

ls

Expected output:

main.py
requirements.txt
models
nodes

6. Install Dependencies

Install Python dependencies:

pip install -r requirements.txt

If using PyTorch:

pip install torch torchvision torchaudio

Installation may take several minutes.

7. Download Models

Create model directories:

mkdir -p models/checkpoints
mkdir -p models/vae
mkdir -p models/loras
mkdir -p models/controlnet

Upload or download Stable Diffusion models.

Example structure:

models
 ├ checkpoints
 │   └ sd_xl_base.safetensors
 ├ vae
 ├ loras
 └ controlnet

8. Start ComfyUI

Run the server:

python main.py --listen --port 8188

You should see logs like:

Starting ComfyUI server
Listening on port 8188

9. Access the ComfyUI Interface

CNB typically provides public ports.

Open:

http://workspace-url:8188

Example:

https://workspace-abc123.cnb.dev:8188

You will see the ComfyUI visual workflow editor.

10. Running ComfyUI in the Background

To keep ComfyUI running after closing the terminal:

Install screen:

sudo apt install screen

Start screen session:

screen -S comfyui

Run ComfyUI:

python main.py --listen --port 8188

Detach session:

CTRL + A + D

11. Updating ComfyUI

Pull latest updates:

git pull

Reinstall dependencies if necessary:

pip install -r requirements.txt

Restart the server.

12. Troubleshooting

Port Cannot Be Accessed

Check if ComfyUI is running:

ps aux | grep python

Make sure port 8188 is exposed by CNB.

Python Version Error

Check Python version:

python --version

Recommended:

Python 3.10+

Out of Memory

ComfyUI may require large memory for models.

Possible solutions:

• Use smaller models
• Increase workspace RAM
• Enable swap memory

13. Recommended Workflow

Typical workflow on CNB:

Fork repository
      ↓
Import repository into CNB workspace
      ↓
Install dependencies
      ↓
Upload models
      ↓
Run ComfyUI
      ↓
Access via public URL

14. Advantages of Running ComfyUI on CNB

Advantages include:

• No local GPU required
• Cloud-based development
• Easy sharing
• Fast setup
• Persistent workspace

This is especially useful for:

• AI experimentation
• teaching environments
• collaborative workflows

Note: Choose whichever installation method you find most comfortable. If you encounter problems during installation, you can check the troubleshooting section later in this guide or search for solutions online.