Anthropic Releases Self-Hosted Claude Managed Agents With 8-Hour Safety Response Window

Anthropic has launched a self-hosted version of Claude Managed Agents, its platform for building and running AI agents. Customers running the self-hosted option face a requirement to act within 8 hours if the system flags a safety risk. The source excerpt is brief, so details on deployment options and pricing are not confirmed here.

Why it matters: Teams deploying AI agents in their own infrastructure need to plan for the operational obligation of responding to safety flags within a fixed time window.