trycua/cua

Open-source infrastructure for Computer-Use Agents. Sandboxes, SDKs, and benchmarks to train and evaluate AI agents that can control full desktops (macOS, Linux, Windows).

82
/ 100
Verified

This project helps AI developers build, evaluate, and deploy intelligent agents capable of controlling an entire desktop environment, whether macOS, Linux, or Windows. It provides the tools to program agents that can "see" screens, click buttons, and execute code within isolated virtual machines or containers. The output is a robust, tested AI agent that can perform complex computer tasks autonomously. It's ideal for AI researchers and developers working on agentic AI.

13,043 stars and 764 monthly downloads. Actively maintained with 116 commits in the last 30 days. Available on PyPI.

Use this if you are developing AI agents that need to interact with a full desktop operating system to perform tasks, from automating workflows to testing software.

Not ideal if you are an end-user looking for a pre-built automation tool or if you primarily need to build web-based automation without desktop interaction.

AI-agent-development desktop-automation AI-agent-evaluation virtual-desktop-environments AI-agent-training
Maintenance 22 / 25
Adoption 17 / 25
Maturity 25 / 25
Community 18 / 25

How are scores calculated?

Stars

13,043

Forks

805

Language

Python

License

MIT

Last pushed

Mar 13, 2026

Monthly downloads

764

Commits (30d)

116

Dependencies

3

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/agents/trycua/cua"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.