What are the system requirements for ComfyUI?

On Windows, you need an NVIDIA GPU with at least 4GB VRAM for best performance. On Mac, ComfyUI supports Apple Silicon (M1/M2/M3) and Intel Macs, using Metal GPU acceleration. You also need Python 3.11 and Git installed on Mac.

How to Install ComfyUI on Windows and Mac (2026 Guide)

How to Install ComfyUI: Get Started in Minutes

Quick answer: ComfyUI is a free, open-source AI image generation tool. On Windows, download the portable zip from GitHub and run it. On Mac, install Homebrew, Python 3.11, and Git, then clone the repository and run it with a single command.

ComfyUI is a free and powerful tool for creating AI images using Stable Diffusion. This step-by-step guide shows how to install ComfyUI on Windows or Mac — no prior experience required.

What Is ComfyUI?

ComfyUI is a free, open-source program for generating AI images using a visual node-based interface. It is one of the most popular tools for Stable Diffusion because it gives you precise control over every part of the image generation process — from model selection to sampling, upscaling, and more.

ComfyUI node-based interface showing connected workflow nodes for AI image generation — ComfyUI's node-based interface — connect nodes to build powerful AI image generation workflows.

Unlike simple text-to-image tools, ComfyUI lets you build custom workflows, chain multiple models, and automate complex image generation pipelines. It works with Stable Diffusion, FLUX, SDXL, and many other model families.

Why Install ComfyUI Locally?

Running ComfyUI on your own computer gives you full control and unlimited access — no cloud credits, no usage limits, no subscription fees. Here is why local installation is worth it:

Full Control

Runs entirely on your own machine. No data leaves your computer.

Faster Results

GPU-accelerated generation is significantly faster than most cloud tools.

Fully Customizable

Install custom nodes, load any workflow, build your own pipelines.

Any Model

Use any checkpoint model you download — FLUX, SDXL, SD 1.5, and more.

Fast Updates

Access new open-source models as soon as they are released.

Free Forever

No monthly fees. ComfyUI is fully open-source with an active community.

How to Install ComfyUI on Windows

The easiest way to install ComfyUI on Windows is the portable version — no Python setup required. Follow these five steps:

1Download the Portable Version

Go to the ComfyUI GitHub Releases page and click the direct download link for the Windows portable zip file (approximately 2 GB).

ComfyUI GitHub releases page showing the Windows portable zip download link — Download the portable zip from the ComfyUI GitHub releases page.

2Extract the Files

Right-click the ComfyUI_windows_portable_nvidia.7z file and select "Extract All". Place the extracted folder on a drive with enough space (for example, D:\). If your default extractor doesn't support .7z files, download 7-Zip.

3Run ComfyUI

Open the extracted folder.
Double-click run_nvidia_gpu.bat (for NVIDIA GPUs). If you have an AMD GPU, use the AMD batch file instead.
ComfyUI will start and automatically open in your browser.

ComfyUI Windows portable folder structure showing run_nvidia_gpu.bat and other batch files — Open the extracted folder and double-click the correct .bat file for your GPU.

4Load a Workflow and Download a Model

In the ComfyUI interface, click the Template section.
Click Image Generation to load the default workflow.
Download the starter model: v1-5-pruned-emaonly-fp16.safetensors

Place the downloaded model file in this folder:

ComfyUI_windows_portable\ComfyUI\models\checkpoints

ComfyUI models/checkpoints folder showing where to place the downloaded .safetensors model file — Place your downloaded model file in the checkpoints folder.

5Start Creating Images

Refresh ComfyUI in your browser, select your model in the Load Checkpoint node, write a prompt, and click Run. Your first AI-generated image will appear in seconds.

First AI image generated with ComfyUI after installation — example output from the default SD 1.5 workflow — Your first AI-generated image, created with ComfyUI and the SD 1.5 model.

How to Install ComfyUI on Mac

Mac installation uses the Terminal and requires a few one-time setup steps. This works on both Apple Silicon (M1/M2/M3) and Intel Macs. Apple Silicon users get Metal GPU acceleration for faster generation.

Step 1 — Prepare a Folder

Open Finder, go to your Home directory, and create a new folder named AI. This is where ComfyUI will be installed.

Step 2 — Open Terminal and Navigate to the Folder

Open Terminal (Applications → Utilities → Terminal) and navigate to your new folder:

cd ~/AI

Step 3 — Install Homebrew

Homebrew is the package manager that makes installing Python and Git straightforward on Mac. Visit brew.sh and follow the installation instructions. Then verify it worked:

brew doctor

Step 4 — Install Python 3.11

Check if Python is already installed:

python3 --version

If it is not installed or the version is wrong, install it with Homebrew:

brew install python@3.11

Step 5 — Upgrade pip and Setuptools

python3 -m pip install --upgrade setuptools

Step 6 — Install Git

If you don't have Git, download it from git-scm.com/downloads/mac and follow the installer instructions.

Step 7 — Clone ComfyUI

Download the ComfyUI source code from GitHub:

git clone https://github.com/comfyanonymous/ComfyUI.git

Mac Terminal showing git clone command downloading ComfyUI from GitHub — Git clones the ComfyUI repository into your AI folder.

Step 8 — Install ComfyUI Dependencies

Navigate into the ComfyUI folder and install the required Python packages:

cd ComfyUI

pip3 install -r requirements.txt

Step 9 — Launch ComfyUI

Start ComfyUI with:

python3 main.py

If you installed Python 3.11 specifically:

python3.11 main.py

Copy the URL shown in Terminal (usually http://127.0.0.1:8188) into your browser.

Mac Terminal showing ComfyUI server starting and providing the localhost URL to open in a browser — Copy the URL from Terminal and open it in your browser to access ComfyUI.

💡 Tip: Once ComfyUI is running, follow Windows Steps 4 and 5 above to download a model, load a workflow, and generate your first image. The process is identical on Mac.

Starting ComfyUI Again Later

Each time you want to use ComfyUI after the initial setup, open Terminal, navigate to your ComfyUI folder, and run:

python3.11 main.py

Accessing and Using the ComfyUI Web Interface

After launching ComfyUI, open your browser and go to the URL shown in your terminal — usually:

http://127.0.0.1:8188

You will see the node-based ComfyUI interface. From here you can load templates, build custom workflows, select models, write prompts, and generate images. If this is your first time using ComfyUI, read the ComfyUI Interface Overview to learn what each panel does.

Conclusion

ComfyUI is a flexible, free tool for creating AI images on your own computer. Now that your installation is complete, the next steps are to explore different checkpoint models, try custom workflows, and install add-ons via ComfyUI Manager to expand what you can create.

Sample AI image generated with ComfyUI after a successful local installation on Windows or Mac — Sample AI image generated with ComfyUI after installation.

For next steps, explore the ComfyUI Manager installation guide to start adding custom nodes and expanding your workflow options.

Frequently Asked Questions

Yes. ComfyUI is completely free and open-source. There are no subscription fees or usage limits when running it locally on your own computer.

On Windows, you need an NVIDIA GPU with at least 4GB VRAM for best performance (AMD GPUs also work). On Mac, ComfyUI supports both Apple Silicon (M1/M2/M3) and Intel Macs, using Metal GPU acceleration. You also need Python 3.11 and Git installed on Mac.

Download the portable zip from the ComfyUI GitHub releases page, extract it to a drive with plenty of space, double-click run_nvidia_gpu.bat, and ComfyUI will open in your browser. Then download a checkpoint model and place it in the checkpoints folder.

Install Homebrew, then use it to install Python 3.11 and Git. Clone the ComfyUI repository with git clone, navigate into the folder, run pip3 install -r requirements.txt to install dependencies, then launch with python3.11 main.py.

For beginners, v1-5-pruned-emaonly-fp16.safetensors is a great starting model. It is small, fast, and produces solid results at 512×512 resolution. Download it from Hugging Face and place it in ComfyUI/models/checkpoints.

👉 Next: How to Install ComfyUI Manager

Discussion

Join the discussion

💬

No comments yet

Be the first to share your thoughts!