ComfyUI Launcher (BETA)

Run any ComfyUI workflow w/ ZERO setup.

Need help? Join our Discord!

Runs anywhere: - Locally - RunPod - Huggingface Spaces

Features

• Automatically installs custom nodes, missing model files, etc.
• Workflows exported by this tool can be run by anyone with ZERO setup
• Work on multiple ComfyUI workflows at the same time
• Each workflow runs in its own isolated environment
• Prevents your workflows from suddenly breaking when updating custom nodes, ComfyUI, etc.

Demo

Running a workflow json file w/ no setup

https://github.com/ComfyWorkflows/ComfyUI-Launcher/assets/33400216/aa17680d-eee5-4e6d-abc4-9f7551f9a4ad

Requirements

Windows (Windows Subsystem for Linux - WSL) & Linux:

• Docker (w/ GPU support) or Python 3

macOS:

• Python 3

Quick start

Option 1: Docker (Linux & Windows)

Linux

docker run \
--gpus all \ # remove this line if you don't have a GPU or if you're on MacOS
--rm \
--name comfyui_launcher \
-p 4000-4100:4000-4100 \
-v $(pwd)/comfyui_launcher_models:/app/server/models \
-v $(pwd)/comfyui_launcher_projects:/app/server/projects \
-it thecooltechguy/comfyui_launcher

Windows

docker run ^
--gpus all ^ # remove this line if you don't have a GPU
--rm ^
--name comfyui_launcher ^
-p 4000-4100:4000-4100 ^
-v %cd%/comfyui_launcher_models:/app/server/models ^
-v %cd%/comfyui_launcher_projects:/app/server/projects ^
-it thecooltechguy/comfyui_launcher

Open http://localhost:4000 in your browser

Option 2: Manual setup (macOS, Linux, and Windows)

Works for Windows (WSL - Windows Subsystem for Linux), Linux, & macOS

Installation (one-time setup)

git clone https://github.com/ComfyWorkflows/comfyui-launcher
cd comfyui-launcher/

Start ComfyUI Launcher

./run.sh

Open http://localhost:4000 in your browser

If you're facing issues w/ the installation, please make a post in the bugs forum on our discord

Updating

Option 1: Docker

docker pull thecooltechguy/comfyui_launcher

Option 2: Manual setup

git pull

Usage

Using a reverse proxy (advanced)

If you're running ComfyUI Launcher behind a reverse proxy or in an environment where you can only expose a single port to access the Launcher and its workflow projects, you can run the Launcher with PROXY_MODE=true (only available for Docker).

docker run \
--gpus all \ # remove this line if you don't have a GPU or if you're on MacOS
--rm \
--name comfyui_launcher \
-p 4000:80 \
-v $(pwd)/comfyui_launcher_models:/app/server/models \
-v $(pwd)/comfyui_launcher_projects:/app/server/projects \
-e PROXY_MODE=true \
-it thecooltechguy/comfyui_launcher

Once the container is running, all you need to do is expose port 80 to the outside world. This will allow you to access the Launcher and its workflow projects from a single port.

Currently, PROXY_MODE=true only works with Docker, since NGINX is used within the container. If you're running the Launcher manually, you'll need to set up a reverse proxy yourself (see the nginx.conf file for an example).

Using an existing ComfyUI models folder

When starting the ComfyUI Launcher, you can set the MODELS_DIR environment variable to the path of your existing ComfyUI models folder. This will allow you to use the models you've already downloaded. By default, they're stored in ./server/models

Using a different folder to store your Launcher projects

When starting the ComfyUI Launcher, you can set the PROJECTS_DIR environment variable to the path of the folder you'd like to use to store your projects. By default, they're stored in ./server/projects

Donations

If you find our work useful for you, we'd appreciate any donations! Thank you!

Coming soon

• Native Windows support (w/o requiring WSL)
• Better way to manage your workflows locally
• Run workflows w/ Cloud GPUs
• Backup your projects to the cloud
• Run ComfyUI Launcher in the cloud

Credits

• ComfyUI Manager (https://github.com/ltdrdata/ComfyUI-Manager/)
• Used to auto-detect & install custom nodes