Anthropic has introduced a new Managed Projects feature currently being tested on Claude, which allows the AI to autonomously manage tasks and maintain project organization. This feature enhances Claude’s ability to share memory and instructions across sessions, facilitating continuity and efficiency in work streams. It builds on the capabilities that were expanded in July 2026, which included support for scheduled work and cross-device persistence, and may also replace the older Conway system, which is being phased out.

Claude: Claude is Anthropic’s flagship AI assistant optimized for reasoning, coding, and complex project work. Recent platform developments include features that allow sessions to share memory across streams of work and operate in dedicated cloud environments. The spotted Managed Projects feature builds on these foundations to enable more autonomous project management.
Anthropic: Anthropic develops advanced AI systems with a focus on safety and capability. The company builds the Claude series of models along with supporting tools and platforms for agentic workflows. The news reports on internal testing of a Managed Projects feature that extends Claude’s cloud-based capabilities for persistent, autonomous task handling.

Platform Features: Anthropic has expanded Claude’s Cowork capabilities in July 2026 to support background tasks, scheduled work, and cross-device persistence for projects and artifacts.
Agent Infrastructure: Claude Managed Agents entered public beta earlier in 2026, providing dedicated environments for long-running autonomous agents with sandboxing and context management.