Claude Managed Agents Is Here and It Cuts AI Agent Setup From Months to Days

Anthropic has launched Claude Managed Agents, a new product on the Claude Platform that helps teams build and run AI agents at scale without months of complex engineering work.

Building a production-ready AI agent used to require a lot of heavy technical work. Teams had to build their own systems for running agents, handling errors, managing memory, storing files, and keeping everything secure. Claude Managed Agents takes care of all of that automatically.

With this new product, teams simply define what they want the agent to do, what tools it can use, and what kind of environment it should run in. Anthropic handles everything else in the cloud, including security, file storage, monitoring, and long-running task execution that can last hours.

The product is built around four simple ideas. An agent is a reusable setup that includes the model, instructions, and tools. An environment is a secure cloud container where the agent works. A session is when the agent actually runs and completes a task. Events are how users and agents communicate back and forth in real time.

Teams can build and manage agents through the Claude Console web interface, the Claude Code terminal tool, a new command line tool, or directly through the API using Python or TypeScript.

Early users are already seeing results. Companies like Notion, Asana, Rakuten, and Sentry have used it to build specialist agents in under a week and set up infrastructure up to ten times faster than before.

Claude Managed Agents is now available in public beta on the Claude Platform. Some advanced features like multi-agent support and memory are still in research preview.