Anthropic has expanded the capabilities of Claude Code by redesigning its Projects feature to allow users to run multiple AI agents simultaneously within the same project.
The update is designed to make complex development and research workflows easier to manage by allowing different AI agents to work on separate tasks while maintaining access to shared project information.
Under the redesigned Projects system, a single project can contain multiple threads. Each thread can be assigned a different task, allowing several AI agents to work on separate parts of a larger project at the same time.
The agents can share important project resources, including memory, goals, files and generated artifacts. This shared environment allows work completed in one thread to remain connected to the broader project.
Anthropic has also introduced a central coordinator to help organize the different streams of work. The coordinator manages the overall project and directs tasks between the individual threads.
This approach could be particularly useful for projects involving multiple stages or tasks that can be handled independently. Instead of waiting for one AI agent to complete each task sequentially, users can divide the workload across multiple threads.
For example, a software development project could involve one agent working on code, another reviewing changes and a separate agent handling documentation. Because the agents operate within the same project environment, their work can remain connected to shared objectives and resources.
The redesigned system also changes how users can approach larger tasks in Claude Code. Rather than treating every AI interaction as an isolated conversation, multiple threads can now operate as parts of a single coordinated project.
Shared memory and files are particularly important in this setup because they allow individual agents to work with common project context. This can reduce the need to repeatedly provide the same information when moving between different tasks.
The central coordination system is intended to keep these parallel activities organized while ensuring that individual threads remain aligned with the project’s broader goals.
Anthropic’s latest Claude Code update reflects the growing use of multi-agent workflows, where several AI systems or instances can divide a complex assignment into smaller tasks and work on them concurrently.
For developers and other users working on large projects, the ability to run multiple agents in parallel could provide a more structured way to manage coding, research, analysis, file handling and other AI-assisted tasks within one shared workspace.
