Whether you are a Windows, Mac, or Linux user, we will help you to install Comfy UI on your PC. ComfyUI is the Web User Interface for running Stable Diffusion which can be managed locally or remotely on any machine. After releasing of this WebUI, it gained huge popularity amongst AI Artists.
Well, ComfyUI gives you a flow-to-model architecture that helps us to understand how the different modules (nodes) are connected with one another which is beginner-friendly. In simple words, it's easy to understand.
Table Of Contents:
Install Comfy UI on Windows
7. After installation, go to the installed folder where you have installed the ComfyUI select the “run_nvidia_gpu.bat” file to run ComfyUI into your web browser.
8. Now, you also need to install ComfyUI Manager that will help you in installing and managing custom nodes. To install this, download the batch file and save into installed "ComfyUI_windows_portable" folder.
Apart from this, we have a detailed tutorial of understanding various ComfyUI nodes that will give you a clear picture of each function.
Downloading models and checkpoints
 |
| Hugging Face platform |
 |
| CivitAI platform |
1. Now, we have to download any standard Stable Diffusion models either from CivitAi or Hugging Face for image generation.
There are multiple Stable Diffusion Models you can try, we have listed below:
- Install Flux (Dev/Schnell/Pro) (Text to Image model)
- Install HiDream (Text to Image model)
- Install Qwen Image (Text to Image model)
- Install Wan (Text/Image to Video model)
- Install Wan 2.2 (Improved variant of Wan)
- Install Stable Diffusion 3.5 (Text-to-Image model)
- Install Frame Pack (Text/Image-to-Video model)
- Install Stable Video Diffusion (Image-to-Video model)
- Installing AnimateDiff with SDXL (Image-to-Animation model)
- Installing AnimateDiff with SD1.5 (Image-to-Animation model)
2. All the models usually goes into the "ComfyUI/models/diffusion_models" folder. If you are using the GGUF model variant then the models gets stored into your ComfyUI/models/unet directory.
4. Enjoy and start your image art generation with ComfyUI.
Install ComfyUI on Mac
Now, we want to mention some important facts about Mac is that the diffusion based models are capable of running and well supported with NVIDIA based systems.
So, you do not get the fastest generation while running the models. Some custom nodes or extensions may not work properly because they were developed and tested mainly on CUDA. But, Mac with Intel based CPUs can handle however it will slow.
1. First, download and install GIT by navigating to its official page.
2. Now, to download from the ComfyUI Repository.
Now just open your terminal and use the command to download from the Git repository:
git clone https://github.com/comfyanonymous/ComfyUI.git
3. After downloading it, just extract the folder using the 7Zip or WinRar application into whatever folder you want to keep in the files.
Keep patience if it takes time to download because the model is very larger in size.
4. Create an installation folder where the Comfy UI files are stored. In our case, we made a new folder named “AI”.
5. Now, open your terminal of Mac by pressing the shortcut key: Command + Space Bar and type terminal.
6. Type into the terminal: cd press once space, then drag and drop the created folder into your terminal like we have shown in the above image. This will just add the path to your created folder.
7. Now, we have to install an application named “Homebrew”. This is the package manager that helps the users to install and run open-source software. Install it by going to their official page.
8. Just copy the code for installation by
clicking the COPY button. Move back
to your terminal paste the code and press Enter. If it requests for a password just type your own Mac password and
press Enter (twice). This installs
all required dependencies like pip, python, etc.
9. Now, type brew doctor and press Enter.
If this gives you some error, you will see a
Warning message just follow the steps and just copy and paste that command and
press Enter.
Again
type brew doctor and you will get
the message that Your system is ready to
brew.
10. Now to check what Python version has got, type
python3 –version
and press Enter. If you want to upgrade the Python version if Comfy UI is not supported then type
brew
install python@3.11
Here, in our case, we are installing Python
3.11 which is the latest version. Place the latest version in place of 3.11 to get
the new version installed.
If you are getting error for python version not supported then its recommended to install Python 3.10.6
Now, install PIP as well. Type this command and press enter :
pip3
install –upgrade pip setuptools
Now, install Pytorch. Go to pytorch official page copy the code for installation and paste it into Terminal and press Enter. At last, we have successfully installed all the dependencies to run ComfyUI.
11. Now, just open your terminal and type command to clone Comfyui repository provided below:
git clone https://github.com/comfyanonymous/ComfyUI.git
This command helps to download from the Git repository from official Comfyui.
12. All the files have been downloaded to our newly created folder “AI”. Type on terminal:
cd and
drag and drop the ComfyUI folder to the terminal. This copies the file path to
the terminal. Press Enter.
Again type
pip3
install -r requirements.txt
This will download and install all the libraries and files.
Now you can open the Comfy UI application which runs stable diffusion by typing:
python3 main.py
A new URL will be provided called a local host (Starts from https://127.0.0.1……). Just copy and paste on into your browser, and ComfyUi will be opened to your browser.
If you have turned off the Mac and want to open ComfyUI then just move into the ComfyUI home directory, open your terminal into that folder and type "python3 main.py" and you will get the dashboard in just a few seconds.
13. There are multiple Stable Diffusion Models we have which you can follow:
- Install Flux (Dev/Schnell/Pro) (Text to Image model)
- Install HiDream (Text to Image model)
- Install Qwen Image (Text to Image model)
- Install Wan (Text/Image to Video model)
- Install Wan 2.2 (Improved variant of Wan)
- Install Stable Diffusion 3.5 (Text-to-Image model)
- Install Frame Pack (Text/Image-to-Video model)
- Install Stable Video Diffusion (Image-to-Video model)
- Installing AnimateDiff with SDXL (Image-to-Animation model)
- Installing AnimateDiff with SD1.5 (Image-to-Animation model)
Install Comfy UI on Linux
1. Make sure you have installed the Git installed on your computer. If not yet, then you can go to Git official website to download and install the current version of Git on your Linux-based system.
2.Install and Setup Python 3.10.6 or greater that's actually supported by ComfyUI. During the installation just check Add Python to Path option.
2. You just need to clone the repo then just use this command:
git clone https://github.com/comfyanonymous/ComfyUI.git
And download from the Git repository.
3. After downloading it, just extract the folder into whatever folder you want to keep the files. To unzip the zipped folder using 7ZIP or WinRAR.
Note: Keep patience if it takes time because the model is very larger in size.
4. There are multiple Stable Diffusion Models you can install by navigating to our tutorials:
- Install Flux (Dev/Schnell/Pro) (Text to Image model)
- Install HiDream (Text to Image model)
- Install Qwen Image (Text to Image model)
- Install Wan (Text/Image to Video model)
- Install Wan 2.2 (Improved variant of Wan)
- Install Stable Diffusion 3.5 (Text-to-Image model)
- Install Frame Pack (Text/Image-to-Video model)
- Install Stable Video Diffusion (Image-to-Video model)
- Installing AnimateDiff with SDXL (Image-to-Animation model)
- Installing AnimateDiff with SD1.5 (Image-to-Animation model)
){kind=link}