AI NewsDev toolsReported

Anthropic redesigned Claude Code Projects to split goals across parallel threads

The New Stack reports that Anthropic is rolling out a redesigned Claude Code Projects in beta on Pro and Max plans, with a coordinator that breaks an engineering goal into parallel threads, shared memory across them, and the warning that each thread counts as a full Claude Code session against usage limits.

AI News

Editorial3 min read

LinkedInX
Anthropic redesigned Claude Code Projects to split goals across parallel threads

Why it mattersA team that switches to Projects gets automatic work splitting and shared context across sessions, and the plan limit is reached sooner because every running thread counts as its own Claude Code session.

The New Stack reports on 17 September that Anthropic is redesigning Claude Code Projects around a coordinator that takes an engineering goal, breaks it into smaller jobs, and hands them to multiple Claude Code sessions running in parallel. Anthropic said Pro and Max subscribers get the redesigned Projects in beta from Thursday, through cloud sessions in Claude Code, with access expanding over the coming week.

What the coordinator does

An Anthropic representative told The New Stack that "Claude scopes the request, delegates the work, coordinates parallel threads, reviews the outputs, and assembles the finished result." Once a developer sets a goal, the coordinator decides how to split the work across threads. Each thread runs as its own Claude Code cloud session, on its own branch and its own copy of the repository, and can use subagents, loops, and workflows on smaller pieces of the job.

A developer can watch the whole Project from the main chat, or open an individual thread to inspect it or redirect it, without having to start and manage each Claude Code session by hand. Two threads that end up changing the same code are handled as a merge conflict, like any other pull request.

What each Project remembers

Anthropic is adding shared memory across threads inside a Project. Information one thread learns can carry into the others without a developer supplying the same context again. The New Stack lists a changed release date, the reason a feature was dropped, and the approvers on a particular service as the kinds of thing a Project can retain, with that information carrying forward over days or weeks as work continues.

A user can also track Project-specific usage, choose a model for the coordinator and a model for the worker threads, and pick an effort level for each. The New Stack says support for threads running locally next to a developer's tools and code, including resources behind a private network, is coming soon.

When the plan limit is reached

The New Stack reports that Anthropic says Projects will hit usage limits sooner when multiple threads are running, because each one counts as a full Claude Code session. That is the trade-off a subscriber gets in one sentence: the coordinator replaces manual work-splitting, and the usage limit on Pro or Max is reached faster when threads run in parallel.

Context on the usage limits: The New Stack notes that Anthropic is currently facing a class-action lawsuit, filed last week, from developers on the Max plan who allege that advertised usage increases came with weekly limits which the company did not disclose at first. The paragraph above about hitting limits sooner is one Anthropic wrote itself, and it is the answer a Pro or Max subscriber needs before turning Projects on.

What is released and when

Beta access starts on 17 September for select Pro and Max subscribers, expanding to the rest of those plans over the coming week. Mobile support is described as coming soon. Team and Enterprise access is planned for later. Existing subscribers already using Projects stay on the current version until Anthropic upgrades them, so no team is silently forced onto the new behaviour.

Source

Reporting: Anthropic's new Claude Code feature could drain your plan before lunch, by Amanda Caswell at The New Stack, 17 September 2026.

Reported byThe New Stack

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.

Share
LinkedInX
Start a project