Anthropic has quietly shipped one of the most practical AI releases of the month: a revamped version of Claude Code Projects that lets you manage a whole team of AI agents in the cloud instead of chatting with one assistant at a time. If you've been following the agent wave — like when we wrote about Zed's Delta agents replacing pull requests — you'll know this is where the entire industry is heading. Anthropic just made it considerably more concrete.
The updates keep coming, too — Claude Code finally supports AGENTS.md, which rounds out the multi-agent story nicely.
01.Claude Code Projects: what actually launched
The idea is simple to describe and big in consequence. A Project in Claude Code now runs multiple agents that all coordinate with each other — sharing one memory, a common set of goals, and a shared library of files and artifacts. Each project contains "threads" running different tasks in parallel, and above them sits a "coordinator" that directs everything.
You can talk to each thread individually when you want to nudge one specific task, or monitor and steer everything through the main project chat. It's less "chatbot" and more "project manager with a staff that never sleeps."
02.Branches, merge conflicts and why the design is clever
Under the hood, each thread is a Claude Code cloud session working on its own branch and its own copy of the repository. The coordinator keeps the work organised, but if two threads touch the same code, the overlap is resolved as a merge conflict — exactly like any pull request. Any thread can also split its own delegated work into smaller pieces using subagents, loops and workflows when a large assignment needs to finish faster.
Honestly, what surprised me most is how unglamorous the design is — in the best way. Anthropic didn't invent a new "agent protocol"; they leaned on the version-control workflow every developer already knows. If you've ever resolved a Git conflict, you already understand how to supervise this thing.
03.Availability: who gets it, and when
At launch, threads run in the cloud, and Anthropic says support for local tools and code is coming "very soon." The beta is available starting today for select Claude Pro and Max subscribers, with access expanding to all Pro, Max, Team and Enterprise users later — and eventually to Cowork and regular Claude chats as well.
No India-specific pricing was announced, but Pro plans have historically been the affordable entry point for Indian students and freelancers, and the Pro tier is where this lands first. If you're already on Pro, keep an eye on your account over the next few weeks.
04.What this changes for developers and freelancers
Here's the thing nobody's talking about: until now, running multiple agents meant juggling tabs, copy-pasting context between windows, and hoping nothing collided. That overhead is exactly what killed most DIY "agent team" setups. A shared memory plus automatic merge-conflict resolution removes both problems in one stroke — the agents literally pass files to each other through the project's shared library.
For a one-person freelance shop in India, this is the shape of a whole new service model: one developer, five parallel agents, and a delivery pipeline that used to need a small team. The bottleneck shifts from writing code to reviewing and directing it — which is precisely the skill gap worth closing this year.
If you want to build that skill on a structured path instead of trial and error, our GPT-6 Astra — Zero to Hero Guide has a full chapter on AI agents (plus one on cost optimization, because five parallel cloud agents can quietly burn through a monthly budget if you're not careful).
Key Takeaways
- Claude Code Projects now supports multiple AI agents with shared memory, goals and files
- Each thread is a cloud session on its own repo branch; a coordinator agent directs the work
- Code overlaps resolve as standard merge conflicts, just like pull requests
- Beta is live for select Pro and Max users; Team, Enterprise, Cowork and wider rollout follow
- Local tools and code support is "coming very soon" per Anthropic
Frequently Asked Questions
What is new in Claude Code Projects?
Anthropic has relaunched Projects inside Claude Code so you can run multiple AI agents under one roof with shared memory, goals and files.
How do the parallel agents work together?
Parallel “threads” each work on their own branch and copy of the repo, while a coordinator agent keeps everything organised. Overlapping work is resolved as a normal merge conflict, like any pull request.
Who can try it?
The beta is live for select Claude Pro and Max subscribers, with a wider rollout and local support coming.
What This Means for You
If you're a student or early-career developer in India, this is your cue to stop treating AI as a autocomplete-on-steroids and start thinking in terms of directing teams of agents — because that's the job description that's emerging. For freelancers and small studios, parallel agents plus merge-conflict safety rails mean you can credibly take on bigger projects without hiring — though at that point the bottleneck is usually the client pipeline, not delivery, and our LinkedIn Client Kit covers that side with copy-paste pitch scripts and a client-outreach system. And the fastest way to get from "I use ChatGPT sometimes" to "I run agent workflows" is structured practice — the GPT-6 Astra — Zero to Hero Guide walks you from prompting fundamentals all the way to building your first agents, with cost control and safety baked in.