100100 Ideas AI
Submit Tool

cua

MCPOpen source

Open-source drivers, cloud fleets, and benchmarks for building and evaluating computer-use AI agents.

28kMIT

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

Pros

  • +Cross-OS desktop automation drivers for macOS, Windows, and Linux
  • +Isolated cloud desktop fleets plus local macOS/Linux VMs via Lume
  • +Built-in benchmarks and specialized CUA-S1 decision models for evaluation

Cons

  • −Cloud fleet pools can retain paid capacity after a claim ends, requiring careful cleanup
  • −Background delivery and full automation depend on platform-specific support boundaries
  • −Broad surface area (drivers, VMs, SDKs, benchmarks) implies a steep learning curve

Target audience: AI engineers and researchers building, training, or evaluating computer-use agents that operate real desktop environments.

Related tools

Other open-source tools that share tags with cua.

headroom

MCPApache-2.0

Open-source context compression layer for AI agents, cutting tokens 60-95% for JSON and 15-20% for coding.

67k·LLM Ops·Open source