Anthropic’s Claude Code and Cowork can control your computer
TL;DR
Anthropic has added computer-use capabilities to Claude Code and Cowork: the AI can autonomously open files, control browsers, and run dev tools with no setup required.
Key Points
- The feature is available as a research preview exclusively for Claude Pro and Max subscribers and is currently limited to macOS.
- Claude asks for permission before taking autonomous actions but can continue working even when the user steps away from the computer.
- The feature builds on computer-use capabilities introduced with Claude 3.5 Sonnet in 2024, now extended to the Code and Cowork environments.
Nauti's Take
Anthropic is doing what OpenAI tried with Operator, but baked directly into developer tools – that is the smart play. Anyone already using Claude Code gets Computer Use without extra setup or a separate interface.
The permission prompt sounds reassuring, but in practice it is a hurdle most users will click away once and forget. The real test comes when the feature expands beyond macOS to Windows and Linux – only then will it become clear whether the architecture is robust enough for broad deployment.
Sources