Computer control for artificial intelligence agents
CUA provides an open source infrastructure for computer-capable artificial intelligence agents. It brings together sandbox, software development kit (SDK) and benchmark tools under one roof for the purpose of training and evaluating agents that can control desktop operating systems.
Updates
- November 8, 2026: Stars 31,903 → 31,961, latest release cua-spaces-v0.22.3 (November 7, 2026).
- November 7, 2026: Stars 31,693 → 31,903, latest release cua-spaces-v0.22.0 (November 7, 2026).
- November 4, 2026: Stars 31,612 → 31,693, latest release cua-spaces-v0.20.0 (November 3, 2026).
- November 3, 2026: Stars 31,550 → 31,612, latest release cua-spaces-v0.19.4 (November 3, 2026).
What you get
- Control desktop apps in the background
- Isolated sandboxes for different operating systems
- Benchmarking tools to measure agent performance
Installation
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/libs/cua-driver/scripts/install.sh)"pip install cuaRunning it
# Install Lume
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/trycua/cua/main/libs/lume/scripts/install.sh)"
# Pull & start a macOS VM
lume run macos-sequoia-vanilla:latestIf you don't write code
I want to develop a computer usage agent using the CUA infrastructure. Help me set up the basic Python structure that will allow my agent to interact with desktop applications in the background, make mouse clicks, and send keyboard input. Create a sample code sketch that runs commands and takes screenshots in a Linux environment using the CUA Sandbox SDK.
Repository: https://github.com/trycua/cua. Follow the official instructions in the README of this repository for setup and execution.
Before starting, check which OS we are on and if the necessary tools (e.g., Python, Node.js, Git, Docker) are installed. Install any missing ones step-by-step according to this system.
Create a new, empty folder to work in. Do not touch any files outside this folder and do not delete anything. Tell me what you are going to do in one sentence before running each command.
Tell me what I should see on the screen at the end of each step. If an error occurs, I will paste the error message to you exactly as it is, and we will fix it together.
Related dictionary terms
Links
TreScout did not build this tool · we found it in GitHub trends and wrote it up. This page describes the repository as of 2026-06-16: The star count and our text belong to that day, the repository may have changed since. Check the repository link for the current state. The introductory text was prepared with AI from the repository's own documentation, and this page was machine-translated from the Turkish original · the Turkish version prevails.