Anthropic Cuts AI Agent Dev Time to Weeks

Anthropic has launched Claude Managed Agents, a cloud service designed to accelerate AI agent development from months to weeks. The platform automates key infrastructure tasks including container setup, state management, tool orchestration, and error recovery, accessible via APIs at eight cents per agent runtime hour plus model usage fees. The service lets developers define tasks, tools, and security rules while agents run in isolated containers. Two research preview features allow agents to spawn sub-agents and auto-refine prompt quality, the latter improving task success by up to 10 points. Early adopters include Notion, Rakuten, and Asana.