<aside> 💡
ComfyUI Desktop (previously known as V1) for Windows & Mac is now available in open beta!
<aside>
Windows (NVIDIA) NSIS x64: Download
macOS ARM: Download
</aside>
<aside>
2a. Install in a new directory
Select a directory for your ComfyUI files. A Python environment will be installed in the selected location and takes around 15GB. System files will also be created during installation as below.
Next, check out some desktop only features like model downloading.
Windows
Mac
</aside>
<aside>
2b. Import from Existing ComfyUI Directory
If you are new to ComfyUI, skip this part.
If you have an existing ComfyUI install, and you can import existing models, user settings, and workflows. Select the root of your old directory and choose the “ComfyUI
” folder (e.g., D:\\ComfyUI_windows_portable\\ComfyUI
). This step works for both Windows and Mac.
Custom nodes cannot be imported yet.
Select the “ComfyUI” folder to import resource
</aside>
https://youtu.be/VviaJ85mPQg?feature=shared
<aside>
The Desktop application stores logs, models, and other files on your system. You can go to theHelp
menu to open the folders conveniently.
</aside>
<aside>
4a. Templates
We provide template workflows under the “workflow” side bar to help you get started. For new users to start your first generation, click Templates
under the Workflows
sidebar. Choose any template workflow to enter the workflow window.
4b. Auto Model Download
ComfyUI Desktop can download models and place them in the correct directory, if they are defined in the workflow json. All of the template workflows have models embedded.
Click the blue button to open Templates
4 basic templates provided
Follow the dialog to download the models needed
https://youtu.be/Eg1JW5NehOM?feature=shared
</aside>
<aside>
5a. ComfyUI Manager
ComfyUI Manager is now shipped by Default. You can now install nodes directly from the ComfyUI registry with access to the latest, semantically versioned nodes. You can search for a custom node in Manager and choose to display them with frequency. Check out more preference options for custom nodes in Settings
!
</aside>