Renowned AI company Anthropic has recently introduced a major upgrade to its developer tool Claude Code, officially adding an integrated browser window. This new feature enables Claude to open, read, click, and type on external web pages directly within the application.
With this upgrade, developers can avoid the hassle of frequently switching between applications, allowing the AI to efficiently retrieve technical documentation, track issue records, and access other critical resources directly within the tool. The built-in browser uses a convenient tab-based management architecture, and users can quickly activate it with simple keyboard shortcuts.
Security and Control Combined
To prevent unexpected unauthorized behavior from the AI, Anthropic has deployed multiple strict security review mechanisms at the technical level. A classifier within the system rigorously screens all writing operations on external websites, ensuring that the core control remains firmly in the user's hands.
Without explicit user authorization, Claude will not be able to perform any actions such as purchasing goods, creating new accounts, or bypassing CAPTCHAs. Additionally, the browser always runs in a completely clean isolated environment configuration and does not store any history login credentials.
Flexible Organization Management
In response to security concerns from enterprise users, organizations can freely restrict the range of external websites that the AI can access by setting up a whitelist. If an enterprise has more advanced compliance requirements, management can even choose to completely disable this browser tool.
For users who wish for Claude to perform actions within their logged-in sessions, the official recommends using a dedicated Chrome browser extension. This combination not only demonstrates the rapid advancement of AI agent (Agent) technology but also strengthens the security measures for the practical implementation of the tool.
