Anthropic Opens Self-Hosted Claude Managed Agents Private Beta for Enterprises

According to ChainCatcher, Anthropic has opened a private beta of self-hosted Claude Managed Agents, allowing companies to run the agent framework, execution sandbox, access credentials, and session history on their own cloud servers or in-house data centers, rather than having Anthropic host them. In May, Anthropic began allowing companies to self-host sandboxes, while Anthropic continued managing the systems that orchestrate agents, call models, and store session records. The new option brings those control systems into the company’s own environment. Companies can limit which servers agents access, manage API keys, and use OpenTelemetry to record and review agent actions, including during code repair, software testing, and system troubleshooting. Claude models will still run in the cloud, so prompts, code, and tool results must be sent to the model provider. Companies can access Claude through Anthropic, Amazon, Google, or Microsoft cloud services, or route all model requests through their own servers to control access, providers, and costs. The beta is limited to approved applicants, who must contact Anthropic’s customer team for deployment files.