Your AI agents can now control your Google Home devices
TechCrunch Sarah Perez ● Covered by 2 sources
Google Home is opening its devices to AI agents through MCP. That means an agent can read history and control gadgets after you sign in and approve it.
Based on reporting by TechCrunch, Sarah Perez — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Google is letting AI agents reach into Google Home, and it’s starting with early access to its Model Context Protocol server. The pitch is simple: if an agent supports MCP, it can work with a user’s smart home devices and event history without the usual custom integration dance.
That opens the door to a very specific kind of home automation. People can ask an agent to review camera summaries, check what’s been happening around the house, control connected devices, and even build custom dashboards in plain language. Google says the setup runs through a Google Cloud project and Home MCP, after which the agent asks the user to sign in and grant access.
The support is broad across the Google Home ecosystem. That includes Nest doorbells and thermostats, plus “Works with Google Home” devices such as light bulbs, including Matter gear. Google has already used MCP in other parts of its business, from Cloud and data platforms to developer tools and Google Workspace. Home, though, is aimed squarely at consumers trying out AI agents for everyday chores.
For now, the rollout is limited. Access starts today and will continue over the coming weeks for U.S. subscribers on Google Home Premium Advanced, the $20-a-month plan that also includes longer event-based video history, descriptive notifications, detailed alerts, video search tools, and daily summaries. Google wouldn’t say when, or if, the feature will reach other plans or markets. It is asking early users for feedback through its Smart Home for Developers Community.
My take — AI-written commentary, not fact-checked reporting
This is the kind of feature that makes AI agents feel less like demos and more like infrastructure. The catch is obvious: once software can talk to your lights, cameras, and thermostat, permission prompts stop being paperwork and start being the product. Google knows it, which is why the paywall comes first and the broad rollout comes later.
Read more about this at: TechCrunch