Claude Code relaunches Projects to manage multiple AI agents in the cloud
The Verge Stevie Bonifield ● Covered by 3 sources
Claude Code brought back Projects to run several AI agents together in the cloud. They share memory and goals, and clashes turn into regular merge conflicts.
Based on reporting by The Verge, Stevie Bonifield — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Claude Code’s revived Projects feature is built for people who want more than one agent working at once. Instead of treating each run as a one-off chat, a project gives the agents shared memory, shared goals, and access to the same library of files and artifacts.
The setup is organized around threads. Each thread handles a different task in parallel, while a coordinator keeps the whole thing from turning into a pile of loose ends. It’s a little like other tools that manage groups of AI agents, including Grok Bot.
Under the hood, each thread is its own Claude Code cloud session. It works on a separate branch and its own copy of the repo, which keeps the agents from stepping on each other too early. When two threads do touch the same code, the conflict is handled the old-fashioned way: as a merge conflict, just like any other pull request.
And the threads can branch out again from there, splitting delegated work further if needed. The pitch is simple enough: use the cloud to juggle multiple agents without pretending they’re all reading the same notebook at the same time.
My take — AI-written commentary, not fact-checked reporting
This is the right way to sell agent tools: fewer magic claims, more plumbing. Shared memory, branches, merge conflicts — boring words, which is usually a good sign. The AI industry could use a lot less “autonomous brilliance” and a lot more version control with a sense of shame.
Read more about this at: The Verge