Skip to content
Get packs

Guide · Open source

How to use Open-Generative-AI: install, run and generate images and videos

Updated · written with AI from the project's own documentation and checked against our catalogue data

Open-Generative-AI (Anil-matcha/Open-Generative-AI) is a desktop and web studio for creating AI images and videos from one interface. It gives access to a large set of image and video models and is released under the MIT licence, so the code is free to inspect and modify. This guide shows you how to get it running, what to expect on each platform and what its limits are.

Open Generative AIAnil-matcha/Open-Generative-AI · ★ 32k · MIT

Quick start

  1. 1Pick hosted or desktop. Use the browser version if you want no setup. Choose the desktop app if you want a local install.
  2. 2Download the installer. Get the DMG, EXE, AppImage or .deb for your system from the releases page.
  3. 3Clear the macOS block. Run the xattr command on the app in /Applications, or use the System Settings Open Anyway option.
  4. 4Pass the Windows SmartScreen prompt. Click More info, then Run anyway, since the installer is not code-signed.
  5. 5Make the Linux AppImage runnable. Run chmod +x on the AppImage, then start it. Install libfuse2 if it fails on an older system.
  6. 6Sign in and generate. Create a free account in the hosted version or open the desktop app and choose a studio to start generating.

What Open-Generative-AI is and who it suits

The project describes itself as a free, open-source alternative to AI video platforms, offering 600+ models across 14 studios, including Image, Video, Audio, Lip Sync, Cinema and Workflows. The README says it has no content filters and that you can also use a hosted version in the browser.

It fits people who want to try many image and video models in one place, creators who want a free, self-hostable front end, and developers who want to read or modify an Electron and JavaScript app. It is less suited to anyone who needs fully offline generation, since the README describes generation through hosted model services.

Choose between the hosted version and the desktop app

The fastest route is the hosted version at muapi.ai. The README says you can use all studios in the browser with no Node.js or setup, and that you need a free account to start generating.

For a local app, the README provides one-click desktop installers for macOS (Apple Silicon and Intel), Windows (x64) and Linux (Ubuntu x64). It says no Node.js or terminal is required for these installers.

  • Hosted: open the muapi.ai link in your browser and sign up for a free account.
  • Desktop: download the installer for your platform from the releases page.

Install the app on macOS

Because the app is not notarised by Apple, Gatekeeper blocks it on first launch. Drag the app to /Applications, then run the command below in Terminal. You only need to do this once.

If you prefer not to use Terminal, try opening the app, then go to System Settings, Privacy & Security, and click Open Anyway, then Open.

  • Mount the DMG and drag the app to /Applications.
  • Open the app once, then right-click it and choose Open, then Open again on the dialog.
xattr -cr "/Applications/Open Generative AI.app"

Install the app on Windows

Windows SmartScreen may warn you because the installer is not code-signed. Click More info, then Run anyway. The app installs to %LocalAppData% with a Start Menu shortcut.

Install or build the app on Linux

Linux users can run the AppImage or install the .deb package. The README also lets you build the installers yourself from source. Generated files are written to the release/ folder.

If the AppImage fails to start on an older system, install libfuse2. On Ubuntu 24.04 and later, an AppArmor policy can block Chromium, which the README covers in its own section.

  • AppImage: make it executable, then run it.
  • Debian/Ubuntu .deb: install it with apt.
git clone https://github.com/Anil-matcha/Open-Generative-AI npm run electron:build:linux chmod +x "release/Open Generative AI-*.AppImage" ./release/Open\ Generative\ AI-*.AppImage sudo apt install ./release/open-generative-ai_*_amd64.deb sudo apt install libfuse2

Get help and understand the costs

The README links a Discord server for discussion and support. The project is MIT licensed, so the code itself is free to use and change.

Not every model is free. The README mentions paid plans and credits for the related hosted service, and it lists subscription ranges for other consumer platforms as a comparison. Check the current pricing page for any service before you rely on it.

  • Licence: MIT.
  • Support: the Discord link in the README.
  • Comparable projects to look at: ComfyUI, Stable Diffusion WebUI, InvokeAI, Pinokio.

What people use it for

  • Compare outputs from several image models in one interface.
  • Create short AI videos with the Video studio.
  • Build a self-hosted front end for generative media you control.
  • Study and modify an Electron and JavaScript studio app.
  • Use the desktop app on a machine where you do not want to manage Node.js.

Frequently asked questions

Is Open-Generative-AI free?

The code is MIT licensed, so you can use, inspect and modify it freely. The README notes that the related hosted service uses paid plans and credits, so not every model or generation is free.

Does Open-Generative-AI need Node.js to install?

The desktop installers are described as requiring no Node.js or terminal. Building from source uses npm, so you need Node.js for that route.

Why does macOS say Open Generative AI can't be opened?

The README says the app is not notarised by Apple, so Gatekeeper blocks it on first launch. Run the xattr command on the app in /Applications or allow it under System Settings, Privacy & Security.

Can Open-Generative-AI run fully offline?

No. The README says generation goes through hosted model services and does not say which parts run locally, so do not rely on it for offline work.

Does Open-Generative-AI have content filters?

The README advertises no content filters. If you need moderation or safety filtering, you should not assume this tool provides it.

What are alternatives to Open-Generative-AI?

The README lists ComfyUI, Stable Diffusion WebUI, InvokeAI and Pinokio as comparable projects. Compare them on the model services and local features you need.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Open-Generative-AI suits people who want many image and video models in one free, MIT-licensed front end, and developers who want to read the code. It is not a good fit if you need offline generation, guaranteed moderation, or every model at no cost. If those matter, look at the comparable projects listed in the README.

Install options & downloads