SwarmUI is a web interface for generating images and video with models on your own GPU. It supports Stable Diffusion, Flux and other image models, video models such as Wan and LTX-2, and some audio models. The Generate tab is a form with prompts, sizes and sliders that a beginner can use straight away; the Comfy Workflow tab exposes the full ComfyUI node graph underneath for people who want it.
It earns its place by sitting between the two camps. ComfyUI is powerful but intimidating, and simpler interfaces run out of road quickly. SwarmUI installs and runs ComfyUI as its backend for you, then adds conveniences on top: an image editor, a grid generator for comparing settings, model downloads, and the ability to spread work across several GPUs or machines, which is where the name comes from.
The limits: it is still labelled beta, it needs a decent NVIDIA or AMD GPU or an Apple Silicon Mac, and the installer downloads ComfyUI, Python packages and models, which takes a while and a lot of disk space.
- Repository: github.com/mcmonkeyprojects/SwarmUI
- Licence: MIT (MIT License)
- Language: C#. Stars: 4,707. Forks: 477. Last push: Oct 21, 2026.
- Scan: safe, Oct 26, 2026, commit d74f8ce
Who it is for
People with a capable graphics card who want to make AI images or video locally, especially those who found ComfyUI too much to start with but want room to grow into it.
Getting started
1. Linux: download and run the installer (needs git and Python 3.11 or 3.12)
wget https://github.com/mcmonkeyprojects/SwarmUI/releases/download/0.6.5-Beta/install-linux.sh -O install-linux.sh && chmod +x install-linux.sh && ./install-linux.sh2. macOS on Apple Silicon (needs dotnet and Python from Homebrew)
git clone https://github.com/mcmonkeyprojects/SwarmUI && cd SwarmUI && ./launch-macos.shOn Windows, download install-windows.bat from the latest release page and run it from a folder outside Program Files. When the browser opens, follow the install page; it downloads the backend and a starter model.
Safety scan
We cloned mcmonkeyprojects/SwarmUI at commit d74f8ce on Oct 26, 2026 and ran the checks described on the GitHub Tools page: credential patterns, decode-and-execute code, install-time scripts, committed binaries, risky CI workflows, every host the code talks to, known vulnerabilities in pinned dependencies, and project hygiene. A person read every hit. This is what we found.
- No secrets, no suspicious code patterns and no bare-IP URLs in about 107,000 lines of C#, JavaScript and configuration.
- Three committed Windows binaries in launchtools/7z/win (7za.exe, 7za.dll, 7zxa.dll) are the standalone 7-Zip, which the README discloses under Legal and uses to unpack downloads on Windows. They are not built from this repository, so you are trusting the copy that was committed.
- launchtools/install-linux.sh is 27 lines: it clones the repository from GitHub and starts the launcher, without sudo. The first-run installer then sets up ComfyUI and pip packages such as rembg and opencv; some optional features pull in AGPL code (ultralytics for face detection).
- Model downloads come from huggingface.co and civitai.com, which is where the network hosts in the code point.
- No lockfile was found for the C# or Python parts, so no dependency advisories were queried. Two workflows, no pull_request_target, no third-party actions. No security policy, Dependabot or CodeQL.
What the scanner counted
| Check | Result |
|---|---|
| Secrets | None found. |
| Suspicious code | None found. |
| Install-time code | 2 installer scripts |
| Committed binaries | 3 executable or compiled objects committed; listed under the raw findings. |
| CI workflows | 2 workflows. None use pull_request_target. No third-party actions. |
| Network hosts | 29 distinct hosts referenced from source; most often huggingface.co, github.com, civitai.com, civitai.red. No URLs to bare IP addresses. |
| Known vulnerabilities | No lockfile to check: dependencies are declared as ranges, so what gets installed is whatever is current on the day. |
| Project hygiene | Has licence file, contributing guide. Missing security policy, automated dependency updates, CodeQL. |
| OpenSSF Scorecard | Not scored: the project is not in Scorecard's weekly index. |
The raw findings
Every hit the scanner wrote out, with a link to the exact line at the scanned commit. Secrets candidates are redacted.
Installer scripts (2)
- launchtools/install-linux.sh, 27 lines; talks to github.com
- launchtools/run_tests.sh, 11 lines
Committed binaries (3)
launchtools/7z/win/7za.exe: PE (Windows executable), 821 KBlaunchtools/7z/win/7za.dll: PE (Windows executable), 284 KBlaunchtools/7z/win/7zxa.dll: PE (Windows executable), 162 KB
By the numbers
| Stars | 4,707 |
|---|---|
| Forks | 477 |
| Contributors | 77 |
| Commits | 5,043 |
| Open issues | 151 |
| Open pull requests | 3 |
| Releases | 10 |
| Latest release | 0.9.8-Beta |
| Licence | MIT |
| Main language | C# |
| Project age | 2 years |
| Last push | Oct 21, 2026 |
| Tracked files | 479 |
| Lines of code | 106.6K |
| Checkout size | 34 MB |
Lines by language: C# 42.8K, JavaScript 31K, JSON 11.4K, Markdown 8,989, CSS 6,711, Python 4,527.
Questions
Is SwarmUI free?
Yes. SwarmUI is MIT-licensed and the author has said it will stay free with no paywall. Models you download have their own licences.
What is the difference between SwarmUI and ComfyUI?
SwarmUI uses ComfyUI as its engine. ComfyUI is a node graph you wire up yourself; SwarmUI adds a simpler form-based Generate tab and extra tools, while still letting you open the raw ComfyUI graph in its own tab.
Does SwarmUI work on a Mac?
Yes, on Macs with Apple Silicon (M1 or later). Intel Macs are not supported.
This post is part of GitHub Tools, where every repository is cloned and scanned before it is written up. The scan is a snapshot of one commit on one day; the repository has moved on since, so check it before you install.