Anthropic has launched the public beta of Claude Managed Agents, a service that lets businesses build and deploy AI agents on Anthropic's own cloud infrastructure. Instead of spending months building sandboxed execution environments, credential management, checkpointing, and tracing systems, developers can define agents via natural language or YAML and run them on Anthropic's platform. Pricing is based on standard API token costs plus $0.08 per active session-hour and $10 per 1,000 web searches. Some features — including advanced memory tooling, multi-agent orchestration, and self-evaluation loops — remain in limited research preview. The launch marks Anthropic's move beyond model provision into the infrastructure layer for enterprise agent deployment.
Sort: