Claude Managed Agents Can Now Run Inside Your Own Security Perimeter

At a glance

Anthropic has launched a private beta of self-hosted Claude Managed Agents, which lets companies run the whole service in their own environment. Developers can build agents that work on internal code and systems and run them on company infrastructure.

Key points

1. 1

   The harness, sandboxes, credentials, and session history all run in the company's own cloud or on-premise servers.

2. 2

   Model requests can go directly to Anthropic, through a cloud provider such as Amazon Bedrock, or through a company gateway.

3. 3

   Credentials stay outside the sandbox, and each environment can be limited to an allowed list of hosts.

4. 4

   Every step and tool call is recorded and can be exported to existing monitoring tools.

Why it matters

Companies that want agents working inside their systems, with no one checking every step, get direct control over their data and access.