Claude Managed Agents

Claude Managed Agents is a service where Anthropic runs long-duration AI tasks for you on its own servers. You don't need to sit in front of a computer watching the whole time — Anthropic manages the job so it keeps running for hours if needed.

Think of it like hiring a contract worker for your company. You set a budget upfront for how much this worker can spend, and it stops itself once that limit is reached. When it hits a tricky decision, it can be set up to check with a more capable "supervisor" model before proceeding. If you've organized your company documents in advance, the worker can also look them up on its own when needed.

This is different from tools where a developer chats directly with an AI to write code on their own machine. That kind of tool works alongside a person in real time. Claude Managed Agents, by contrast, is more of a backend execution service — it keeps running on its own, within rules you've set, even while you're away.

Articles mention it like: "On August 9, a hard budget cap was set for Claude Managed Agents..." It's easy to confuse this with Claude Code's "auto-resume" feature from around the same time. Claude Code is a tool developers use directly while watching the screen, whereas Claude Managed Agents is a separate service that manages the long-running agent session itself in the background — through budgets, advisor models, and similar controls.