On October 21, Anthropic announced the launch of web-based Claude Code, upgrading AI programming from a command-line tool to a cloud-based visual interface. Developers can now delegate programming tasks directly in their browser, completing code modifications, testing, and pull request generation without opening a terminal.
A New Visual Programming Experience
Web-based Claude Code bridges the gap from the terminal to the browser. Developers simply connect their GitHub repository and describe their requirements, and Claude will automatically execute programming tasks in a cloud environment hosted by Anthropic. All operations run in an isolated sandbox, supporting real-time progress tracking and interactive adjustments, allowing developers to participate in and guide the AI’s programming process throughout.
Parallel Processing and Security
Claude Code supports running multiple project tasks in parallel, significantly improving development efficiency. Developers can handle routine fixes, backend changes, and more simultaneously, with the system automatically generating clear change summaries and pull requests.
In terms of security, it uses an isolated sandbox execution mechanism, with all Git operations completed through a secure proxy. Claude can only access authorized repositories, ensuring code security. Developers can also customize network whitelists to perform operations like dependency downloads while maintaining security.
Anthropic has also launched a mobile preview version, enabling developers to manage coding tasks on iOS devices, extending programming work to a mobile environment. This feature is currently available to Claude Pro and Max users as a preview.