Google is letting AI agents take over your smart home. Starting today, any digital assistant that supports the Model Context Protocol (MCP) can now control your Google Home devices, review camera summaries, and monitor activity across your connected hardware. The update opens a natural-language channel between your voice commands and your home’s appliances, and it arrives through an early-access rollout that will expand to subscribers over the coming weeks.
The change is a significant shift for anyone who owns a smart home and wants a conversational interface that speaks directly to their hardware. Instead of opening a dedicated app or pressing a button, you could simply tell an AI agent what you want done, and the agent would relay the command to your lights, locks, and cameras.
What MCP Does
MCP is a server protocol that allows AI agents to securely connect to smart home devices and access their event history. When an agent supports MCP, it can authenticate with a device, retrieve information about recent events, and send instructions to perform actions. The protocol handles authentication and authorization, so the agent only sees what it is allowed to see.
Google is rolling out early access to its MCP server for the Google Home ecosystem. That means any agent that speaks MCP — including Claude, Hermes, OpenClaw, ChatGPT, and Google’s own Antigravity — can now work with your connected devices. The list covers a wide range of agents, from independent projects to established platforms, which suggests this is not a closed club. Anyone building an AI assistant can join if they implement the protocol.
The system supports any device in the Google Home ecosystem. That includes Google Nest doorbells and thermostats, as well as “Works with Google Home” and Matter devices like light bulbs. The coverage is broad enough that most common smart home setups should work with the new connection.
How the Connection Gets Set Up
Setting up the connection requires a few steps. First, you need to create a Google Cloud project and configure it to use Home MCP. Then you provide the MCP configuration details to your chosen agent and ask it to set up the link. The agent will then ask you to sign in and grant permissions.
A setup guide will also be offered in the Google Home Developer Center. That is the official documentation where developers can find the technical details. For consumers, the process is abstract, but the guide exists for anyone who wants to dig deeper.
The whole flow is designed to be handled through the agent itself. You tell the agent what you want to connect, and it walks you through the rest. That is the point of the protocol: it should make the underlying machinery invisible to the user.
What You Can Actually Do With It
The update gives you several ways to interact with your smart home. You can use natural language instructions to:
- Review camera summaries
- Monitor smart home activity
- Control your connected devices
- Build your own custom smart home dashboards
The dashboard feature is worth noting. Rather than relying on a fixed interface provided by Google, you can ask your agent to surface the information you care about in a format you specify. The source describes this as building your own custom smart home dashboards, which implies a degree of flexibility beyond simple command-and-control.
The camera summaries are another practical feature. Instead of scrolling through a feed, you can ask your agent to review the day’s footage and report back what it saw. The protocol’s event-history capabilities make this possible, since the agent can request recorded events from the device and present them in natural language.
Who Gets Early Access
The rollout starts today and continues over the coming weeks to subscribers who pay for Google Home Premium Advanced in the U.S. That is the more expensive, $20-per-month subscription tier, which provides features like longer event-based video history, descriptive notifications and detailed alerts, tools to search video history, daily summaries, and more.
Google wouldn’t comment on whether or when the MCP would roll out more broadly, such as to other subscription tiers or markets. That is a deliberate hedge. The company is releasing the protocol to a subset of its customer base, and it is not committing to a wider expansion yet.
Feedback From Early Adopters
Google is soliciting feedback from early adopters during this period by way of its Smart Home for Developers Community. The community is where the company will gather input on how the protocol performs in the wild.
The feedback loop is standard practice for new releases. Google wants to know whether the MCP works reliably, whether it meets the expectations of its users, and whether there are edge cases that need fixing. The community is the designated channel for that conversation.
Why Google Is Doing This
Google already supports MCP in other areas of its business, including its Google Cloud and data platforms, developer tools, and Google Workspace. The protocol is not new to the company, though the source does not state how long it has been used internally. The Home ecosystem is a consumer-facing application of the technology.
The move is targeted more directly at consumers who are experimenting with AI agents designed to handle everyday tasks. That framing suggests the company sees a market for people who want a single conversational interface for their home automation. Rather than juggling multiple apps, you could ask one agent to manage everything.
What the Rollout Means for You
If you subscribe to Google Home Premium Advanced, you can try this today. Create a Google Cloud project, configure it for Home MCP, and point your agent at it. The setup guide in the Developer Center will walk you through the specifics.
The experience depends on your agent. The list of supported agents includes both established platforms and smaller projects, so the quality of implementation will vary.
The benefits are tangible. You can control your devices with natural language, review camera summaries, and build custom dashboards. The MCP protocol makes those interactions secure and reliable, since the agent authenticates with each device before sending commands.
What We Don’t Know Yet
The rollout is limited to the premium subscription tier in the U.S., so if you live elsewhere or you are on a different plan, you won’t see this immediately. Google hasn’t announced a timeline for broader availability, and it declined to comment on whether the protocol will expand to other tiers or markets.
The protocol itself is open to any agent that implements it, so the ecosystem could grow quickly once developers start building support. The list of existing agents shows that major platforms are already on board, which reduces friction for users who already rely on those services.
Final Word
Google’s MCP rollout is a meaningful step forward for smart home integration. It gives AI agents a standardized way to control your devices and access your event history, which removes the need for proprietary apps and siloed interfaces. The protocol is open, the agents are numerous, and the rollout is underway.
The rollout started today. If you have the subscription, you can try it now. If you don’t, you will need to wait for a broader announcement.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.