Anthropic launches web-based Claude Code with parallel tasking for developers

Anthropic has launched a web version of its AI coding assistant, Claude Code, allowing developers to create and manage coding agents directly in their browser. This release marks a major upgrade from the original command-line interface, now offering access to both Pro and Max plan subscribers. Users can access it through the Claude website under a new “Code” tab or try a preview version in the Claude iOS app for mobile workflows.

Developers can connect Claude Code to their GitHub repositories, allowing the agent to read issues, make commits, and execute instructions through a terminal-like interface. It now supports multiple tasks in parallel, removing the single-task limit from the CLI version and enabling bulk operations like fixing multiple bugs or updating backend routes simultaneously.

These coding tasks run in isolated sandboxes with strict restrictions on network and filesystem access. Git interactions pass through a secured proxy, ensuring credentials and code remain protected. During execution, Claude Code gives real-time progress updates, and users can interactively guide or modify agent behavior with additional prompts. Early reports suggest the cloud-based system is faster and more capable than its CLI predecessor, though detailed documentation is still forthcoming.