Claude Code introduces revamped Projects feature for managing multiple AI agents in the cloud

Claude has introduced a revamped “projects” feature in its Claude Code platform that lets developers run several AI agents within a single workspace. The new design provides a shared memory, common goals, and a unified library of files and artifacts, enabling multiple agents to collaborate on a codebase without needing separate environments. Each project contains parallel “threads,” each operating as an independent Claude Code cloud session on its own branch and repository copy, while a central “coordinator” oversees task distribution and progress tracking.

The coordinator functions similarly to tools such as Grok Bot, orchestrating work across threads and handling any overlap in code modifications. When two threads edit the same portion of code, the system generates a merge conflict that is resolved through the standard pull‑request workflow, ensuring consistency and version control. This architecture allows threads to further subdivide their assigned tasks, creating a scalable, organized approach to multi‑agent development. The update aims to streamline collaborative AI‑driven coding and is detailed in a full report by The Verge.