Anthropic rebuilds Claude Code projects around parallel cloud threads

Anthropic has rebuilt projects in Claude Code so that one request fans out into a team of cloud sessions directed by a coordinator. In a blog post on September 17, the company said the redesigned projects are in beta for select Claude Pro and Max subscribers. Claude now “scopes the request, delegates the work, coordinates parallel threads, reviews the outputs, and assembles the finished result,” and it keeps working after you step away from your computer.

The mechanism is threads. Each thread is a full Claude Code cloud session working on its own branch and its own copy of the repository, and the coordinator routes work into new or existing threads. If two threads touch the same code, Anthropic says the overlap is resolved as a merge conflict, just like any other pull request. A thread can split its own work further using subagents, loops and workflows.

Anthropic’s own examples show the kind of job it’s built for. Set a goal of cutting your app’s checkout p75 latency, and Claude profiles each endpoint, tests optimisations and opens pull requests in parallel threads. Connect your API, web and mobile repos with a goal of retiring a deprecated v1 endpoint, and it creates a thread per repo, migrates the callers, runs the tests and tells you which PRs need to merge first. That’s a step beyond the single-session loop we described in how coding agents actually work.

Memory is the second piece. Every thread adds to and draws from a shared project memory, which Anthropic says reduces the need for complex prompt engineering, and a library collects the files you add alongside the artifacts Claude produces. You can also tune the coordinator: how often it checks in, how often it starts new threads, and how detailed each update is. ZDNet’s David Gewirtz noted that the feature doesn’t appear to be strictly autonomous, since threads begin when you hit the Start button for each thread Claude recommends.

Running work in parallel is the easy part to sell. The part that actually earns trust is that every thread still lands as a pull request someone reviews. That’s the difference between a tool platform teams can adopt on their own terms and one they have to fight to keep visible.

Mitch Ashley, vice president and practice lead, The Futurum Group, via DevOps.com

The catch is cost. Because each thread is a full Claude Code session and a project can run several at once, Anthropic warns that projects can reach usage limits faster. Its answer is project-specific usage tracking, plus the option to pick the model and effort level for the coordinator chat and for the worker threads separately.

The other limit is where the work runs. Threads live in the cloud today, and Anthropic says running on your own machine, alongside local tools and code and behind your network, is “coming very soon.”

That matters for anyone who runs Claude Code in a terminal on local files. Gewirtz, who works that way, said the launch rules them out for now. The Verge compared the setup to Grok Bot and other tools that manage groups of AI agents, and reported that access is due to expand to Cowork and regular Claude chats as well as Team and Enterprise plans.

Who	When, per Anthropic
Select Pro and Max subscribers using cloud sessions with no existing projects on web or desktop	Beta from September 17
More Claude Code users on Pro and Max	Over the coming week
All of Claude, plus Team and Enterprise plans	After that
Existing projects on Pro and Max	Unchanged for now, upgraded as the rollout reaches chat and Cowork
Threads on your own machine	“Coming very soon”

Anthropic isn’t the only lab betting on this shape. Meta made a similar bet on parallel sub-agents with Muse Code, and the question DevOps.com raised applies to both: whether the logic that decides what to delegate, when to merge and what counts as done holds up on a codebase messier than a demo. The wider Pro and Max rollout over the coming week is the first chance to find out, and the local version is the one to watch after that.