Anthropic launches Projects to turn Claude into a team workspace

If you are tired of repasting the same brand guidelines, code snippets, and background context every time you open a new chat window, Anthropic just fixed your biggest pet peeve.

The AI lab rolled out Projects for Claude.ai Pro and Team subscribers, giving users dedicated workspaces to organize their chats around specific goals or tasks.

A 500-page book in every workspace

Each Project comes packed with Claude's massive 200K token context window. Anthropic says that is roughly the equivalent of a 500-page book.

Users can upload codebases, internal research, interview transcripts, and style guides directly into a project. Powered by the newly launched Claude 3.5 Sonnet, Claude can ground its answers in your internal files instead of hallucinating in a vacuum.

You can also set custom instructions per project. That means you can tell Claude to answer strictly as a senior security engineer in one folder, while acting as a marketing copywriter in another.

Live previews and shared team feeds

Projects works seamlessly alongside Anthropic's new Artifacts feature. Whenever Claude generates frontend code, diagrams, or documents, they pop up in a dedicated side panel where users can view and edit in real time.

For Team plan subscribers, Anthropic added a shared activity feed. Teammates can post snapshots of their best prompts and workflows, letting everyone learn from the top Claude power users in the company.

Consulting firm North Highland noted that hundreds of its consultants use Claude to review complex 10-K filings and craft proposals, speeding up analysis tasks by up to 5x.

Anthropic also stressed that no project data or chat logs will be used to train its generative models without explicit user consent. Integrations with popular external enterprise tools are slated for the coming months.

Why it matters

We think Anthropic is executing a brilliant product shift here. While OpenAI built Custom GPTs like mini-apps meant for public sharing, Anthropic is designing Claude as a collaborative internal knowledge hub.

For creators and developers, this makes Claude.ai immediately more useful. Instead of wrestling with external vector databases or messy pastebins, you can drop your entire codebase into a project, test code via Artifacts on the fly, and keep your workspace organized. If you already pay for Pro, you should move your active workflows into Projects right away.

The strategic move puts serious pressure on ChatGPT Team. By pairing a 200K context window with live Artifacts previews, Anthropic makes long-form drafting and web coding feel cohesive.

The only real catch is scalability. Dumping everything into a 200K context window works well for mid-sized projects, but it does not fully replace dedicated RAG pipelines for massive enterprise repositories. How well Anthropic delivers on its promised third-party integrations will determine whether Projects becomes a true workflow backbone or just a glorified folder system.