Skip to content

[GitHub Trending] trycua/cua

7.9 relevance
Score Breakdown
technical depth
8
novelty
9
actionability
7
community
6
strategic
8
personal
9

Scored daily by a customisable AI persona to surface the most relevant engineering leadership news.

Open-source infrastructure for computer-use agents, highly novel and relevant to AI agents.

AI/ML github.com
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). - trycua/cua
Summary

Cua is an open-source toolkit for building, benchmarking, and deploying computer-use agents across macOS, Windows, and Linux. It provides background drivers that let agents click and type without stealing focus, a unified sandbox API for any OS or VM, and a benchmark suite (Cua-Bench) for evaluating agents on tasks like OSWorld and ScreenSpot. The project also includes Lume, a macOS VM manager leveraging Apple's Virtualization.Framework for near-native performance on Apple Silicon.

Author

trycua