Claude Code’s revised projects adds AI orchestration, but local developers must wait

Anthropic has announced plans to revamp the concept of projects in Claude Code, introducing new features that offer significant improvements for certain work types. These new capabilities include AI orchestration, superpowers, and hard limits. Currently, this functionality is available to select Claude Pro and Max subscribers using cloud sessions in Claude Code, provided they do not have existing projects on the web and desktop.

The revamped projects feature allows Claude to scope requests, delegate work, coordinate parallel threads, review outputs, and assemble finished results. It also enables users to steer progress throughout the process and maintain functionality even after closing their laptop. One of the primary challenges in launching multiple agents is ensuring they work together effectively.

With Anthropic's new projects, developers can manage this by creating addressable threads, each functioning as a full cloud session with its own branch and copy of the repository.

Each thread can be named and steered mid-flight, allowing developers to dig into a running thread and redirect it as needed without wasting time or consuming tokens. Additionally, threads can be aware of each other, with Claude learning more about project details and applying them to its work. This shared memory reduces the need for complex prompt engineering and allows work to span multiple repositories.

Anthropic also plans to introduce mobile check-ins and improve thread persistence, enabling Claude to run even when the computer is shut down. While this new feature does not currently support local developers, Anthropic has stated that running threads alongside local tools and code on the user's machine is forthcoming. The terminal and CLI functionalities remain unaffected, and users will still have access to these tools.