Anthropic launches Claude Code Projects to enhance developer productivity

Anthropic today launched Claude Code Projects, a new feature that lets developers keep an ongoing conversation with Claude that remembers past work and can delegate tasks. The update was announced at the company’s headquarters in San Francisco and is available to all Claude 3 users on the platform’s API and web interface. The move positions Anthropic as a competitor to GitHub Copilot and OpenAI’s Codex, offering a more persistent, project‑aware assistant.

The need for “always‑on” coding help has grown as developers turn to AI for routine tasks. Existing tools can answer a question or generate a snippet, but they lose context after a session ends. In 2023, more than 15 million developers used some form of AI coding assistant, yet many reported frustration when the tool forgot earlier parts of a project. Claude Code Projects addresses that gap by storing a project’s history in a dedicated workspace. The assistant can now keep track of files, dependencies, and design decisions, and it can hand off subtasks to other AI agents or human teammates.

Key details of the new feature include a project dashboard, a “memory” toggle, and a delegation engine that can assign work to other AI agents or external services. Anthropic says the system can remember up to 10,000 turns of dialogue, which covers most medium‑sized projects. The team behind Claude has built the feature on the same safety‑first principles that guided the original model, adding filters to prevent hallucinations in code. Early beta testers report that the assistant can generate boilerplate code, review pull requests, and even draft documentation without losing context after hours of work.

What happens next is a race to prove the value of persistent AI in software teams. Anthropic plans to open the feature to public beta in the next month and will integrate with popular IDEs such as VS Code and JetBrains. If developers adopt it, the line between human and AI collaboration will blur further, and companies may start treating AI as a full‑time team member. The launch signals that AI assistants are moving beyond one‑off prompts toward long‑term partnership, a shift that could reshape how code is written, reviewed, and maintained.