Anthropic's Claude Can Now Operate a Mac in Place of the User

Anthropic has given Claude the ability to control a Mac in real time: the AI takes over routine actions in the interface in place of the user.

What Claude can do now

In Claude Cowork and Claude Code, the model has gained access to full control of macOS. It sees the screen, clicks, opens files and interacts with applications the way a human would.

• Opening and editing files
• Controlling the browser (navigating, filling in forms)
• Launching development tools
• Carrying out multi-step tasks through the system interface
• Delegating tasks from a smartphone via the Dispatch feature

How Dispatch works

Dispatch lets users send tasks to Claude straight from their phone. The user describes a task on the smartphone, and the model carries it out on the desktop on its own. This removes the need to sit at the computer to kick off routine operations.

Context: computer-use agents are gaining momentum

Anthropic is following the same path as OpenAI with Operator and Google with Project Mariner. All the major AI labs are now betting on agents that can work inside a real operating system environment, not just answer questions in a chat.

What this means

The line between a chatbot and a full-fledged digital employee is blurring faster than expected. For developers and technical users, this means being able to delegate entire workflows: not individual prompts, but sequences of actions.