Cua
Open-source computer-use agent tools with an MIT license, local VMs on Apple Silicon, and hosted desktop fleets across Linux, Windows, macOS, and Android.
Cua gives AI agents access to computers they can inspect and operate, with tools for desktop automation, local virtual machines, and hosted fleets. It's for developers building agents that work across native apps and browsers, or evaluating how well those agents complete computer tasks. You bring the agent and model.
Cua Driver connects agents to macOS, Windows, and Linux through MCP, a CLI, or SDKs. It works with Claude Code, Codex, Cursor, and OpenClaw. Where the app and platform support it, agents can interact in the background without moving your pointer or taking focus. An agent can combine code, APIs, and graphical interfaces within one task.
For work on your own hardware, Lume creates and manages local macOS and Linux VMs on Apple Silicon using Apple's Virtualization.Framework. Cua Fleets provides isolated cloud desktops, with fleet support spanning Linux, Windows, macOS, and Android. Local sandboxes and hosted Fleets share a Sandbox SDK for running commands, capturing screenshots, and interacting with apps, though their runtime requirements differ. Fleets runs in the cloud; Lume's VMs run locally.
Cua Bench lets developers build tasks, evaluate agents, and export action trajectories for training. Its simulated tasks can run without a VM, Docker, or a model API key. The open-source repository uses the MIT license.
