Google lets AI agents control your smart home via MCP

Google rolled out early access to a Model Context Protocol (MCP) server for its Google Home ecosystem on Wednesday, letting any AI agent that supports MCP — including Claude, Hermes, OpenClaw, ChatGPT, and Google Antigravity — securely work with a user’s smart home devices and access their event history. The update lets people use natural-language instructions to review camera summaries, monitor smart home activity, control connected devices, and build custom smart home dashboards.

To set it up, users create a Google Cloud project, configure it for Home MCP, and hand the configuration details to their chosen agent, which then asks them to sign in and grant permissions; Google is also publishing a setup guide in its Home Developer Center. The system works with any device in the Google Home ecosystem, including Nest doorbells and thermostats and “Works with Google Home” (Matter) devices like light bulbs.

Google already supports MCP elsewhere in its business, including Google Cloud, data platforms, developer tools, and Workspace, but the Google Home integration is aimed more directly at consumers experimenting with AI agents for everyday tasks. Access is rolling out starting today and over the coming weeks to U.S. subscribers of Google Home Premium Advanced, the $20-a-month tier that includes longer event-based video history, descriptive notifications and detailed alerts, video-history search tools, and daily summaries. Google wouldn’t say if or when the MCP would expand to other subscription tiers or markets, but says it’s gathering feedback from early adopters through its Smart Home for Developers Community.

Why it matters

By exposing Google Home to any MCP-compatible agent — including rivals like ChatGPT — Google is betting that opening up its smart home ecosystem to outside AI assistants will drive more value (and Premium Advanced subscriptions) than keeping control locked inside its own apps.