Claude Code relaunches Projects to manage multiple AI agents in the cloud

Running multiple AI coding assistants in parallel within a single project becomes possible, speeding up complex tasks.

Anthropic has relaunched its Claude Code Projects feature, enabling users to manage multiple AI agents within a single cloud-based environment. This revamped feature allows for shared memory, goals, and file libraries across these agents, which operate in parallel "threads." A central "coordinator" directs these threads, with each thread functioning as an independent Claude Code cloud session working on its own copy of a repository. Overlapping code changes between threads are handled like standard merge conflicts in a PR workflow. Threads can further delegate tasks to subagents, loops, and workflows to accelerate large assignments. Users can interact with individual threads or monitor the entire project. While initially cloud-based, local tool and code support is planned soon. The feature is rolling out in beta to select Claude Pro and Max subscribers, with broader access to all Pro, Max, Team, and Enterprise users, as well as Cowork and regular Claude chats, planned for the future.

The relaunch of Claude Code Projects signals a growing trend towards more sophisticated orchestration of AI agents for complex software development. By enabling parallel execution and merge conflict resolution, Anthropic is addressing the challenges of coordinating multiple AI collaborators, moving beyond simple single-agent interactions. This approach allows for the breakdown of large coding assignments into manageable sub-tasks, potentially accelerating development cycles and improving efficiency for teams.

The introduction of shared memory, goals, and artifacts within projects is crucial for maintaining coherence and consistency when multiple AI agents work concurrently. The planned support for local tools and code, alongside the existing cloud execution, will offer developers greater flexibility and control over their AI-assisted workflows. As this technology matures, we can expect increased adoption of such multi-agent systems, fundamentally changing how software is developed and managed by both human and AI contributors.