Anthropic launches Claude Code Projects for persistent developer coordination

Anthropic just handed software engineers something they didn’t know they needed: a version of Claude that doesn’t forget what it was doing the moment a session ends. Claude Code Projects, unveiled on September 17, 2026, is a redesigned coordination layer that lets a single conversation spin up and manage multiple concurrent cloud-based coding sessions simultaneously.

The pitch is simple. Instead of a developer juggling five separate AI coding threads like browser tabs nobody wants to close, Claude acts as the orchestrator, defining engineering goals, delegating work, collecting outputs, and stitching results together.

From storage cabinet to command center

The previous incarnation of Claude Code Projects functioned primarily as a static information repository. The redesign transforms that into a dynamic coordination hub where Claude manages work across parallel cloud threads, including tasks like revising multi-repository projects or working through accumulated bug backlogs, while maintaining context and project state across sessions.

Anthropic introduced web-based agents to Claude Code in October 2025, and cloud sessions followed as the infrastructure layer enabling longer-running interactions. The September 2026 redesign is the point where those pieces combine into a multi-agent management system.

The current rollout is limited to a public beta for select Claude Pro and Max subscribers using cloud sessions. Gradual expansion is expected over the week following launch, with Team and Enterprise access planned for a later date. Developers running local-only workflows are not yet supported, nor are users who already have existing web or desktop projects, who will remain on the prior version until their accounts are upgraded.

What it means to let AI orchestrate your codebase

There’s a catch worth flagging. Running parallel cloud sessions consumes more resources than a single session would, so developers operating under tight usage constraints should expect higher consumption. The local-only exclusion is a more significant limitation in the near term, effectively sidelining developers working in security-sensitive environments or with proprietary code they prefer not to route through cloud infrastructure, with no rollout timeline yet specified for those users.

A broader shift toward agentic AI in engineering

Claude Code itself has evolved substantially since its origins as a command-line interface tool launched around 2024. The trajectory from CLI tool to web-based agent to multi-session orchestration hub tracks the broader industry movement toward AI systems capable of executing multi-step engineering tasks with less human intervention at each step.

Anthropic is competing in a market where GitHub Copilot, Cursor, and a growing roster of AI coding tools are all attempting to move up the value chain from suggestion engine to genuine development partner. Positioning Claude Code Projects as a coordination layer rather than just an assistant puts it in a different category, one where the comparison isn’t just code quality but workflow integration and project management capability.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.