trycua/cua

This page introduces an external open-source repository. It is not an HDATF product.

What it does

Provides desktop automation, isolated cloud desktops, local macOS virtual machines, and benchmarks for computer-use agents. Agents can use commands, screenshots, and application interactions inside these environments.

How it helps ATF

For Harness, its isolated desktops and evaluation benchmarks are worth comparing when designing external execution environments and checking agent task results.

License

MIT Permissive. Commercial use and changes are allowed if the copyright notice is kept.

More in this category

• autonomous-ai/openharness
  A software and hardware platform for running coding agents in persistent terminals across machines. Domain-specific harnesses target circuit-board design, 3D modeling, robot simulation, and game development.
• browser-use/jev-ultrafast
  A browser agent that selects operations and targets from an indexed element table. A small language model generates text only for text-entry operations.
• aipoch/open-science
  A local-first, model-agnostic AI research workbench. Agents read files, search, execute Python and R, and produce research outputs with traceable provenance.
• Human-Agent-Society/reef
  Infrastructure connecting agent inference, feedback, learning and versioned delivery. It supports model-weight training and optimization of prompts, rules and skills.
• ApodexAI/FrontierAgent
  An agent runtime, terminal interface, and evaluation suite for extended research and file work. It supports a stateful ReAct agent and an Agent Team with a coordinator, task board, parallel assignments, and collected reports.

Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.