Anthropic relaunches Claude Code Projects for teams of AI agents

In short: Anthropic has relaunched the Projects feature in Claude Code, letting people manage multiple AI “agents” working on the same coding job in the cloud.

What happened

Anthropic updated a feature called Projects inside Claude Code, its tool that helps with software coding. Projects is designed to run several AI agents at the same time. An AI agent is a version of the AI that can take on a task and work on it, like a helper assigned to a specific job.

In each Project, the work is split into “threads,” which are separate work streams running in parallel (like having several people in a group chat each handling a different to do). A “coordinator” agent oversees the threads and keeps work organized. Users can talk to each thread directly, or use a main project chat to monitor and adjust what is happening.

Anthropic says each thread runs as its own cloud session with its own copy of the code and its own working area. If two threads change the same part of the code, the conflict is handled the normal way developers do it, by resolving a “merge conflict” (when two edits collide and someone must choose what stays).

At launch, these threads run in the cloud, meaning on Anthropic’s servers instead of your computer. Anthropic says support for local tools and local code is coming “very soon.”

Why it matters

For people building software, this is a way to delegate work to several AI helpers at once, instead of asking one AI to do everything in a single conversation. It could make larger coding tasks faster to manage, but it also adds new complexity, since parallel work can clash and still needs human review.

Source: The Verge AI