By Hamza Faisal
Anthropic has opened a private beta that lets companies run its Claude Managed Agents service entirely on their own infrastructure, removing one of the biggest obstacles to letting autonomous AI agents work inside corporate systems.
Announced on November 12, the self-hosted edition moves the whole agent service behind the customer's own security perimeter: the harness that runs the agent loop, the sandboxes where agents execute code, the credentials agents use, and the full history of every agent session. Until now, Managed Agents ran entirely on Anthropic's servers. Anthropic added self-hosted sandboxes in May, which let the code an agent runs execute on customer hardware, but the agent itself and its data still lived with Anthropic. This release brings the rest of it home.
What runs on the customer's side
A self-hosted deployment has three customer-operated pieces. The first is cma-server, which exposes the same Managed Agents API as the hosted service, runs the agent loop, and stores agents, sessions and their histories. The second is a PostgreSQL database, version 16 or later, that holds that history — without it, the server keeps data only in memory and a restart wipes it. The third is the workers that run the shell and file tools inside the customer's sandboxes; the server itself never executes an agent's shell or file tools.
Because the server exposes the same API, existing tooling keeps working: the Claude SDK and Anthropic's ant command-line tool connect to a self-hosted server just as they do to the hosted service, apart from a few documented differences.
The model itself stays put
There is one thing that does not move home: the model. Every request to Claude still goes to an inference provider the customer picks — Anthropic by default, or Amazon Bedrock, Google's agent platform, Microsoft Foundry, or the customer's own model gateway. Self-hosting the service does not by itself keep model traffic inside the customer's network, and teams with strict data rules will need to read the fine print on where their prompts travel.
Anthropic says it has no remote access to a customer's self-hosted server or its logs. The trade-off is operational: customers install each release themselves, and access to the private beta runs through Anthropic's account team, which hands over the release archive. Anthropic has published no pricing, no list of covered countries and no date for general availability.
Not yet a feature-for-feature match
The private beta is not a copy of the hosted service. Scheduled deployments, dreams and webhooks are unavailable on the self-hosted server, and web search and web fetch stay switched off until a team explicitly enables them. Anthropic's documentation describes a Docker Compose setup for development and warns that it is not appropriate for production, pointing teams toward Helm-based deployment tooling for a persistent install.
The pitch is aimed at regulated industries and large engineering organisations that already run agents unattended inside their systems — fixing broken builds, testing releases, triaging alerts and working through bug backlogs — but could not accept credentials and session history sitting on a vendor's servers. For those teams, the question is no longer whether the agent can run inside the perimeter. It is whether running the whole control plane themselves is worth the operational cost.



