trycua/cua

Worth installing(Strong adoption and recent activity)

Scale computer-use 2.0 with open-source drivers, cross-OS fleets, and benchmarks for training, evaluation, and data generation.

Why this repo matters

Provides open-source sandboxes and SDKs for training AI agents to control full desktop environments across macOS, Linux, and Windows.
Includes standardized benchmarks to evaluate computer-use agent performance, addressing critical gaps in current desktop automation testing methodologies.
Demonstrates strong developer adoption with over 21k stars and consistent recent activity in the emerging computer-use infrastructure space.

Key Metrics

Stars: 29k
Forks: 2.1k
Open Issues: 1.1k
Stars (7d delta): 6.9k
Stars (30d delta): 6.9k

Scores & Metadata

Trend Velocity Score: 100.00
Opportunity Score: 89.00
Confidence Score: 100.00
Language: Rust
License: MIT
Topics:
🏷️ agent 🏷️ ai-agent 🏷️ apple 🏷️ computer-use 🏷️ computer-use-agent 🏷️ containerization 🏷️ cua 🏷️ desktop-automation 🏷️ hacktoberfest 🏷️ lume 🏷️ macos 🏷️ manus 🏷️ operator 🏷️ swift

Star History

22.1k stars+3.5k over 34 snapshots6/22/2026 – 9/2/2026

Links & Sources

Related Repositories

codecrafters-io/build-your-own-x

547.5kMarkdown

freeCodeCamp/freeCodeCamp

freeCodeCamp.org's open-source codebase and curriculum. Learn math, programming, and computer science for free.

456.9kTypeScript

affaan-m/ECC

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.

269kJavaScript

multica-ai/andrej-karpathy-skills

A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.

217.5k