Google is opening early access to a new Model Context Protocol (MCP) server for Google Home, allowing AI agents to interact with compatible smart-home devices and access event history.
The update means people can use supported AI agents to perform smart-home tasks through natural-language instructions, including controlling connected devices, reviewing security-camera summaries and monitoring activity around the home.
The early-access rollout began in the United States on Sept. 16, 2026, and is initially limited to subscribers of Google Home Premium Advanced, the company’s $20-per-month plan.
What Google Home MCP Lets AI Agents Do
The new Home MCP connection allows AI agents that support the Model Context Protocol to interact with devices connected to a user’s Google Home ecosystem.
Supported agents can be used to:
- Review summaries from compatible security cameras
- Monitor smart-home activity
- Control connected devices
- Access event history
- Create custom smart-home dashboards
Instead of using separate commands inside a traditional smart-home app, users can give instructions to an AI agent in natural language.
Google says the system is designed to work securely with the user’s connected home devices and account permissions.
Which AI Agents Can Work With Google Home?
Google’s Home MCP server can be used by AI agents that support MCP.
The examples cited in the announcement include Claude, Hermes, OpenClaw, ChatGPT and Google Antigravity.
That means Google Home’s new capability is not limited to Google’s own AI tools. Third-party agents that support the protocol can potentially connect to the smart-home ecosystem after the user completes the required setup and authorization.
How to Connect an AI Agent to Google Home
Setting up the connection requires several steps.
First, users need to create a Google Cloud project and configure it to use Home MCP.
They then provide the MCP configuration information to the AI agent they want to use.
The agent will ask the user to sign in to their Google account and authorize the required permissions.
Google is also expected to publish a step-by-step setup guide through the Google Home Developer Center.
The permissions step is particularly important because the AI agent is being given access to connected devices and information about activity within the home.
Google Nest and Matter Devices Are Supported
Home MCP is designed to work across the Google Home ecosystem.
That includes Google’s own smart-home products, such as Google Nest doorbells and thermostats.
It also supports compatible third-party products that work with Google Home or the Matter smart-home standard.
That can include everyday devices such as smart lights.
The broader device compatibility means the feature is not limited to a single type of smart-home product. Depending on the connected equipment, an AI agent could potentially interact with lighting, climate controls, security devices and other supported hardware.
Google Home Premium Advanced Is Required for Now
The initial rollout is limited to subscribers of Google Home Premium Advanced in the U.S.
The plan costs $20 per month and includes several advanced smart-home features.
Those include longer event-based video history, descriptive notifications, detailed alerts, tools for searching video history and daily summaries.
Google is rolling out Home MCP gradually over the coming weeks.
The company has not said when the feature will become available to other Google Home subscription tiers or users outside the United States.
Why MCP Matters for Smart Homes
The Model Context Protocol is designed to give AI systems a standardized way to interact with external tools and services.
Google already supports MCP in other parts of its ecosystem, including Google Cloud, data platforms, developer tools and Google Workspace.
Home MCP brings that approach into the consumer smart-home environment.
The difference is that an AI agent can now become an interface between the user and physical devices in the home.
A person could potentially ask an agent about recent camera activity, check what happened in the house or interact with compatible devices without manually navigating through several smart-home controls.
Google Is Expanding AI Beyond Chat
The Google Home integration reflects a broader shift in how AI agents are being used.
Instead of simply answering questions or generating text, agents are increasingly being designed to interact with external services and carry out tasks on a user’s behalf.
Smart homes are a particularly direct example because the results can affect physical devices.
That makes authorization and access controls an important part of the Home MCP setup. Users must explicitly sign in and grant permissions before the agent can interact with their Google Home environment.
Wider Availability Has Not Been Announced
For now, Google is keeping Home MCP in early access.
The rollout is focused on Google Home Premium Advanced subscribers in the United States, with access expanding over the coming weeks.
Google has not announced a date for wider availability across other subscription plans or international markets.
As the early-access program expands, feedback from users and the smart-home developer community will help shape how the feature develops.
For Google Home owners, the immediate change is straightforward: supported AI agents can now move beyond conversation and begin interacting directly with the smart-home devices and activity connected to their Google Home ecosystem.
