GitHub

cua

Observability & evals

Observation pool · not formally ranked

This project does not currently meet the main index requirement of two current dimensions across two source types.

— Not ranked

Observed adoptionMissing

—

MomentumCurrent · 2026-09-12

AttentionCurrent · 2026-09-12

Signal confidence Based on how many independent score dimensions currently have data.

Medium2/3 · 1 source types

Methodology v2.0 · snapshot 2026-09-12 · stale after 2 days

About

Build, benchmark, and deploy agents that use computers

Drive native desktop apps in the background. Agents click, type, and verify without stealing the cursor or focus. Use the same CLI and MCP server on macOS, Windows, and Linux from Claude Code, Cursor, Codex, OpenClaw, and custom clients. Linux supports X11 and compositor-specific Wayland routes with explicit limits for raw background input.

Then wire it into Claude Code as an MCP server and your agent can drive the desktop in the background:

Full tool reference, architecture notes, and the optional agent skill pack live here: libs/cua-driver/README.md.

Build agents that see screens, click buttons, and complete tasks autonomously.…

Across sources

GitHub trycua/cua metrics & trend →

22.5k Stars

Stars 22.5k
Forks 1.6k
Commits 4.7k
Releases 676