Google is expanding access to its smart home system by letting third-party AI assistants interact with connected devices and analyze home data through a new system called the Model Context Protocol (MCP). This update allows AI agents like Claude, Open Claw, Hermes, Google Antigravity, and ChatGPT to securely access and control devices in a user’s Google Home setup. The integration, known as Google Home MCP, was introduced by Taylor Lehman, a product manager at Google Home & Nest, who described it as enabling these agents to "work with all of the devices and event history in your Google Home ecosystem." With this new feature, users can give voice commands to review summaries from security cameras, monitor smart home activity, control devices like lights and thermostats, and even create personalized dashboards. To use the integration, users must first set up a Google Cloud project and configure it to use the Home MCP. Then, they need to share the MCP setup details with their chosen AI agent, which will lead them through the sign-in process and permission granting. Google has provided a setup guide in the Google Home Developer Center to help users through the process. The update applies to devices within Google’s ecosystem, including Google Nest products such as doorbells and thermostats, as well as "Works with Google Home" devices, which are compatible through the Matter standard, like smart light bulbs. Access to the MCP is currently limited to users with the Google Home Premium Advanced subscription in the U.S. This subscription costs $20 per month or $200 per year and includes features like extended video history, detailed notifications, and daily summaries of home activity. Google announced that early access to the MCP server for the Google Home ecosystem was launched this week, with continued access for Premium Advanced users in the following weeks. The company has not shared plans for expanding access to other subscription tiers or international markets. The Model Context Protocol was introduced by Anthropic in 2024 and has since been adopted by several major tech companies, including Google, OpenAI, Microsoft, and SpaceXAI. However, some experts have warned about potential security risks, including a lack of strict authentication measures, which have led to data breaches in some instances.