Anthropic launches Claude Code Projects to coordinate persistent development workflows

Anthropic has launched Claude Code Projects, a Claude-powered AI coordination agent designed to handle long-running development work. The tool takes natural language input from developers, breaks tasks into parallel threads, and automatically coordinates them while maintaining shared project memory across sessions.

The new product differs from Anthropic's existing Claude Projects, which is a general workspace for organizing chats and knowledge. Claude Code Projects instead focuses on coordinating agentic software development across multiple Claude Code sessions. Developers can keep talking to a central coordinator, which decides whether each request belongs in an existing thread or a new one. Each worker thread runs as a cloud-based Claude Code session on its own branch and repository copy.

The beta is initially available to select Claude Pro and Max subscribers using cloud sessions in Claude Code, with broader access expected in the week following the September 17 announcement. Cloud threads cannot yet access local files or internal network tools; Anthropic says local support is coming soon.

The launch comes as OpenAI, GitHub and Google also push toward asynchronous and parallel agent workflows. Anthropic's approach places a persistent conversational coordinator above individual sessions and repositories.