According to CoinMeta, Anthropic has announced the opening of a private beta test for Agent managed by Claude. Enterprises can now deploy the entire operating framework, execution sandbox, access credentials, and session history of AI Agent on their own cloud servers or internal data centers, without having to rely on Anthropic for hosting. The new version also moves the core systems responsible for scheduling Agent, calling models, and saving session records onto the enterprise's internal network. Enterprises can control which servers Agent can access, manage API keys, and use OpenTelemetry to record and review every step of the process. However, the Claude models themselves still run in the cloud, and related prompts, code, and tool results still need to be sent to the model service providers. Enterprises have the option to call Claude through Anthropic, Amazon, Google, or Microsoft's cloud services, or they can route all model requests through their own servers to uniformly control access rights, call channels, and fees. This new feature is currently only available for private testing to enterprises that have been approved; to obtain the deployment files, it is necessary to contact Anthropic's customer team.