SwarmUI

SwarmUI v0.9.8 Beta.

Formerly known as StableSwarmUI.

A Modular AI Image Generation Web-User-Interface, with an emphasis on making powertools easily accessible, high performance, and extensibility. Supports AI image models (Stable Diffusion, Z-Image, Flux, Qwen Image, etc.), and AI video models (Wan, Hunyuan Video, etc.), with plans to support eg audio and more in the future.

• Discord Community: Join the Discord to discuss the project, get support, see announcements, etc.
• Announcements: Follow the Feature Announcements Thread for updates on new features.
• General documentation: /docs folder
• Website: SwarmUI.net

---

Status

This project is in Beta status. This means for most tasks, Swarm has excellent tooling available to you, but there is much more planned. Swarm is recommended as an ideal UI for most users, beginners and pros alike. There are still some things to be worked out.

Beginner users will love Swarm's primary Generate tab interface, making it easy to generate anything with a variety of powerful features. Advanced users may favor the Comfy Workflow tab to get the unrestricted raw graph, but will still have reason to come back to the Generate tab for convenience features (image editor, auto-workflow-generation, etc) and powertools (eg Grid Generator).

Those interested in helping push Swarm from Beta to a Full ready-for-anything perfected Release status are welcome to submit PRs (read the Contributing document first), and you can contact us here on GitHub or on Discord. I highly recommend reaching out to ask about plans for a feature before PRing it. There may already be specific plans or even a work in progress.

Key feature targets not yet implemented:

• Better mobile browser support
• full detail "Current Model" display in UI, separate from the model selector (probably as a tab within the batch sidebar?)
• LLM-assisted prompting (there's an extension for it, but LLM control should be natively supported)
• convenient direct-distribution of Swarm as a program (Tauri, Blazor Desktop, or an Electron app?)

Donate

SwarmUI is 100% free and open source forever. If you want to help make sure it keeps pace with the best despite my refusal to paywall access or shove ads down your throat, donate to SwarmUI!

Try It On Google Colab

Google Colab

WARNING: Google Colab does not necessarily allow remote WebUIs, particularly for free accounts, use at your own risk.

Colab link if you want to try Swarm: https://colab.research.google.com/github/mcmonkeyprojects/SwarmUI/blob/master/colab/colab-notebook.ipynb

Run it on a Cloud GPU Provider

Runpod

Runpod template (note: maintained by third party contributor nerdylive123): https://get.runpod.io/swarmui

Vast.ai

Vast.ai template (readme): https://cloud.vast.ai/?ref_id=62897&creator_id=62897&name=SwarmUI

Note it may take several minutes to start up the first time. Check the container logs to see setup progress. Check the template ? info for hints on how to use.

Installing on Windows

Note: if you're on Windows 10, you may need to manually install git and DotNET 8 SDK first. (Windows 11 this is automated).

• Download The Install-Windows.bat file, store it somewhere you want to install at (not Program Files), and run it.
  • It should open a command prompt and install itself.
  • If it closes without going further, try running it again, it sometimes needs to run twice. (TODO: Fix that)
  • It will place an icon on your desktop that you can use to re-launch the server at any time.
  • When the installer completes, it will automatically launch the SwarmUI server, and open a browser window to the install page.
  • Follow the install instructions on the page.
  • After you submit, be patient, some of the install processing take a few minutes (downloading models and etc).

(TODO): Even easier self-contained pre-installer, a .msi or .exe that provides a general install screen and lets you pick folder and all.

Alternate Manual Windows Install

• Install git from https://git-scm.com/download/win
• Install DotNET 8 SDK from https://dotnet.microsoft.com/en-us/download/dotnet/8.0 (Make sure to get the SDK x64 for Windows)
• open a terminal to the folder you want swarm in and run git clone https://github.com/mcmonkeyprojects/SwarmUI
• open the folder and run launch-windows.bat

Installing on Linux

Prereqs

• Install git and python3 via your OS package manager if they are not already installed (make sure to include pip and venv on distros that do not include them in python directly)
  • For example, on some Ubuntu (desktop) versions, sudo apt install git python3-pip python3-venv, or you may need https://launchpad.net/~deadsnakes/+archive/ubuntu/ppa
  • For Debian or Ubuntu Server, sudo apt install git python3-full
  • You'll want Python 3.11 or 3.12. Things should also work fine with 3.10. Do not use 3.13.
  • Make sure python3.11 -m pip --version returns a valid package

Linux Easy Install

• Download the install-linux.sh file, store it somewhere you want to install at, and run it
  • If you like terminals, you can open a terminal to the folder and run the following commands: (Yes this link is still current):
    • wget https://github.com/mcmonkeyprojects/SwarmUI/releases/download/0.6.5-Beta/install-linux.sh -O install-linux.sh
    • chmod +x install-linux.sh
• Run the ./install-linux.sh script, it will install everything for you and eventually open the webpage in your browser.
• Follow the install instructions on-page.

Linux Manual Install

• Install DotNET 8 using the instructions at https://dotnet.microsoft.com/en-us/download/dotnet/8.0 (you need dotnet-sdk-8.0, as that includes all relevant sub-packages)
• Open a shell terminal and cd to a directory you want to install into
• Run shell commands:
  • git clone https://github.com/mcmonkeyprojects/SwarmUI
  • cd SwarmUI
  • ./launch-linux.sh
  • or if running on a headless server, ./launch-linux.sh --launch_mode none --host 0.0.0.0 and/or swap host for cloudflared
• open http://localhost:7801/Install (if it doesn't launch itself)
• Follow the install instructions on-page.

Linux Install Notes

• You can at any time in the future run the launch-linux.sh script to re-launch Swarm.
• If the page doesn't open itself, you can manually open http://localhost:7801

(TODO): Maybe outlink a dedicated document with per-distro details and whatever. Maybe also make a one-click installer for Linux? Can we remove the global python install prereq?

Installing on Mac

Note: You can only run SwarmUI on Mac computers with M-Series Apple silicon processors (eg M1, M2, ...).

• Open Terminal.
• Ensure your brew packages are updated with brew update.
• Verify your brew installation with brew doctor. You should not see any error in the command output.
• Install .NET for macOS: brew install dotnet.
• If you don't have Python, install it: brew install python@3.11 and brew install virtualenv
  • Python 3.11, 3.12, 3.10 are all fine. 3.13 is not, do not use 3.13.
• Change the directory (cd) to the folder where you want to install SwarmUI.
• Clone the SwarmUI GitHub repository: git clone https://github.com/mcmonkeyprojects/SwarmUI.
• cd SwarmUI and run the installation script: ./launch-macos.sh.
• Wait for the web browser to open, and follow the install instructions on-page.

Installing With Docker

See Docs/Docker.md for detailed instructions on using SwarmUI in Docker.

Documentation

See the documentation folder.

Motivations

The "Swarm" name is in reference to the original key function of the UI: enabling a 'swarm' of GPUs to all generate images for the same user at once (especially for large grid generations). This is just the feature that inspired the name and not the end all of what Swarm is.

The overall goal of SwarmUI is to a be full-featured one-stop-shop for all things Stable Diffusion.

See the motivations document for motivations on technical choices.

Legal

This project:

• embeds a copy of 7-zip (LGPL).
• has the ability to auto-install ComfyUI (GPL).
• has the option to use as a backend AUTOMATIC1111/stable-diffusion-webui (AGPL).
• can automatically install christophschuhmann/improved-aesthetic-predictor (Apache2) and yuvalkirstain/PickScore (MIT).
• can automatically install git-for-windows (GPLv2).
• can automatically install MIT/BSD/Apache2/PythonSoftwareFoundation pip packages: spandrel, dill, imageio-ffmpeg, opencv-python-headless, [matplotlib](https://pypi.org/proje