Claude’s Innovative ‘Computer Use’ Feature Empowers AI to autonomously Build, Test, and Debug Applications

Anthropic has launched a pivotal upgrade to its AI assistant, Claude, introducing the “Computer Use” feature to enhance automation and productivity. This feature allows Claude to interact autonomously with desktop applications, replicating human-like operations. Under Dario Amodei’s leadership, Claude can now execute complex workflows—from coding to debugging—without ongoing user input. Integrated within Claude Code, this capability is accessible via command-line on macOS, enabling tasks like opening apps, navigating interfaces, and enhancing software development processes. The setup involves enabling the “computer-use” MCP server by running a command in the terminal, necessitating specific macOS permissions. The feature is currently in research preview for Pro and Max plan users, requiring Claude Code version 2.1.85 or later. With an emphasis on user control and security, each session mandates explicit permission for app access, reinforcing transparency. This upgrade represents a significant shift toward AI as a hands-on productivity tool, potentially transforming user interactions with technology.

Source link