Claude Managed Agents can now run on customer infrastructure in private beta

Anthropic has opened a private beta of self-hosted Claude Managed Agents, a version that lets companies run the agent service inside their own cloud or on-premise environment. Help Net Security dates the launch to November 12. Before this, the only piece customers could self-host was the sandbox where agent code runs, which left credential storage and activity records on Anthropic's side. Access is granted on request through an Anthropic account team rather than through self-serve sign-up.

Under the new setup, the customer runs the Managed Agents server, a PostgreSQL database and the workers that carry out shell and file tools. <cite index="10-7,10-8">Model requests go wherever the customer points them: to Anthropic directly, to Amazon Bedrock, Google's Agent Platform or Microsoft Foundry, or to the customer's gateway. Claude itself still runs outside the customer's network</cite>. <cite index="10-12,10-13">Self-hosted customers can keep those credentials in their own secrets manager. Managed Agents also keeps a log of every step an agent takes</cite>. Anthropic's product page says <cite index="3-13">if your organization has zero data retention in place with Anthropic, it applies to the requests sent to Claude in the same way as any other Messages API call.</cite>

Operating work moves to the buyer, who must patch the server, database and firewall rules and manage capacity. <cite index="2-14">Anthropic limits that protection to MCP tools and "supported API calls," so any other service an agent reaches is worth checking before the agent goes live.</cite> No price or general availability date appeared in the coverage reviewed.