Claude Code Projects is a 📝Claude Code feature in which one ongoing conversation acts as a coordinator, turning a stream of related tasks into parallel worker sessions called threads.
📝Anthropic announced the redesigned Projects on September 17, 2026. As of October 2026 it is in public beta on Pro and Max plans, rolling out gradually, and runs at claude.ai/code, in the desktop app's Code tab, and in the Claude mobile app, but not in the terminal CLI or IDE extensions. A project has three parts: the project conversation, where Claude routes each message to an answer, a new thread, or an existing thread; the threads themselves, usually cloud sessions that each work on their own branch and open pull requests; and the Overview pane, which groups threads as Ready for review, Waiting on you, Working, Landing, Idle, or Resolved.
Context is set once and reaches every new thread. Project instructions (up to 16,000 characters) act as the standing brief, project memory stores decisions and pitfalls Claude is asked to remember, and each repository's 📝CLAUDE.md loads in every cloud thread. Repository skills, .mcp.json servers, hooks, and .claude/settings.json do not load in cloud threads; connectors come from the claude.ai account instead. A thread can run on the user's own computer through Remote Control when a task needs local tools.
Projects suit work that outlasts one session and keeps producing tasks, such as one goal across many repositories or a migration bigger than a session. Coordinator and thread models are set separately, and the enforced limit is 200 new threads per day across a user's projects. For a single task, a plain cloud session fits better; for work that needs the local machine throughout, 📝Claude Code Agent View does.
The role Claude plays in the project conversation is described in 📝Claude Coordinator, and 📝Claude Code Multi-Agent Orchestration compares Projects with every other way to run Claude in parallel.