Give an AI agent a whole computer to operate, in a sandbox
Made by Cua
View profile on GitHub (opens in a new tab)A collection of open-source pieces for letting AI agents use a computer rather than an API: drivers that move the mouse and keyboard, isolated cloud desktops, local macOS virtual machines, small models trained to make computer-use decisions, and a benchmark for scoring how well an agent did.
It is designed to work with whichever agent and model you already have.
You could use it to…
- Give an agent a disposable cloud desktop to work in
- Throw the desktop away when the agent's task is done
- Score how well an agent used the computer with a benchmark
Can I use this?
You'll need A developer setup, and a cloud account for the hosted desktops · Setup needed
Worth knowing This is a toolkit for building on, not a finished app, and the pieces are used separately. Running agents on hosted desktops costs money, and an agent with control of a desktop can do real damage if pointed at the wrong thing — which is why the isolation matters. MIT licence. BuildTube has not run or verified this software.
Why it's here
Rising fast on GitHub this week
- 1,474 people starred it on GitHub this week — 27,658 in total.
Numbers from the snapshot taken 1 October 2026; not refreshed since.
Behind it
See the code on GitHubBuildTube has not run or verified this project. Everything above is written from what the creator published.
Made this?