Self-Hosted Claude Managed Agents (Private Beta)

Run the full Claude Managed Agents service in your own cloud or on-premises, for teams whose code and data cannot leave their environment.

You host the agent harness, sandboxes, credentials, and session history yourself, including the Managed Agents server, its database, and the workers that run shell and file tools.
You choose where requests to Claude go: Anthropic directly, Amazon Bedrock, Google's Agent Platform, Microsoft Foundry, or your own gateway.
Credentials stay out of the sandbox, with a proxy adding them for supported tools. Each environment can be limited to an allowlist of hosts, and every step, including tool calls and their results, is recorded and can be exported through OpenTelemetry.
Private beta for eligible organizations only. Access is requested through your Anthropic account team, and pricing is not published.

Running agents against internal code and systems without moving credentials, files, or session history to a hosted service.

Security rules block hosted agents, or the work touches code that has to stay inside your network, like fixing broken builds or triaging alerts.

Engineering and security teams at larger organizations that already work with an Anthropic account team.

Join the newsletter to get ahead with AI.

I share strategies, new modules, free guides, and tools to get you better at using AI in your work.