Anthropic has relaunched its Projects feature within Claude Code to enable users to run multiple AI agents under a single roof. This update allows the system to maintain shared memory, common goals, and a unified library of files and artifacts across different tasks. Each project now operates with threads that execute distinct jobs in parallel while a central coordinator directs the overall workflow. Under the hood, every thread functions as a separate cloud session working on its own branch and copy of the repository. The coordinator keeps work organised, but if threads modify the same code, the system resolves overlaps as standard merge conflicts just like any other pull request. Each thread can also further split its delegated work to handle complex sub-tasks.
This shift addresses a practical need for managing complex workflows without requiring constant human oversight. By treating parallel threads as independent branches, the tool prevents the chaos often seen when multiple agents edit the same files simultaneously. The approach mirrors established software development practices, ensuring that conflicts are handled predictably rather than causing data loss. It also allows developers to scale operations by delegating specific sub-tasks to individual threads while maintaining a cohesive project structure.
* Threads operate on isolated branches to prevent direct file collisions
* Merge conflicts are resolved automatically using standard repository tools
* The coordinator manages parallel execution without human intervention
