
Anthropic has expanded the reach of its AI coding assistant with the launch of Claude Code on the web, allowing developers to interact with the tool directly from a browser or smartphone. This eliminates the need for terminals or local setups, making it easier to integrate AI-assisted coding into existing workflows. The beta research preview, available to Pro and Max users since October 20, supports connecting a first GitHub repository at claude.com/code.
Claude Code on the web enables developers to delegate multiple coding tasks to the AI, which runs on Anthropic-managed cloud infrastructure. This setup is particularly effective for understanding project structures, mapping repositories, handling bug fixes, performing routine tasks, and managing back-end changes with test-driven verification. By executing tasks in the cloud, developers benefit from scalable, secure processing without overloading local machines.
Each web session runs in an isolated sandbox with network and file restrictions, ensuring security and stability. Real-time progress tracking and Git interactions through a secure proxy help developers maintain visibility and control over the process. Parallel task execution across multiple repositories is supported, and Claude Code can automatically generate pull requests with clear change summaries, streamlining the development workflow.
Anthropic has also extended this web experience to its iOS app, aiming to refine mobile interactions based on user feedback. By offering both browser and mobile access, Claude Code seeks to make AI-assisted development more accessible and flexible, helping developers work efficiently across platforms while maintaining high security and workflow transparency.

