Google is giving AI agents a new way to interact with smart home devices, allowing users to control products connected to Google Home through natural-language commands.
The company has begun early access to a Model Context Protocol (MCP) server for Google Home. The system allows AI agents that support MCP, including Claude, Hermes, OpenClaw, ChatGPT and Google Antigravity, to securely connect with compatible smart home devices and access certain home event history.
Instead of manually opening different apps or navigating through multiple controls, users will be able to communicate with their AI agent and ask it to perform tasks across their connected home. This could include controlling devices, checking activity around the house and reviewing summaries from compatible security cameras.
The integration could also make it possible for users to create customized smart home dashboards through their preferred AI agent. Because the system is based on natural-language interaction, users can describe what they want rather than having to know exactly where a particular setting or device control is located.
For example, an AI agent could help a user review recent camera activity, check what happened at home during a specific period or control compatible lights and other connected devices. The exact capabilities available will depend on the devices connected to Google Home and the permissions granted by the user.
Setting up the connection will require several steps. Users will first need to create a Google Cloud project and configure it to use Home MCP. They will then provide the required MCP configuration information to their chosen AI agent and ask it to establish the connection.
Once the connection process begins, the AI agent will direct the user to sign in and approve the necessary permissions. Google will also provide a setup guide through the Google Home Developer Center to help users configure the system.
The new system is designed to work across the wider Google Home ecosystem rather than being limited to Google's own hardware. Supported products include Google Nest doorbells, thermostats and other compatible devices, as well as third-party products that support "Works with Google Home" or the Matter smart home standard.
That means devices such as compatible smart lights can also be included in an AI-controlled setup, provided they are supported within the Google Home ecosystem.
Google has already adopted MCP across several parts of its business, including Google Cloud, data platforms, developer tools and Google Workspace. Bringing the technology to Google Home takes the same concept into the consumer smart home space, where AI agents could potentially handle more everyday tasks on behalf of users.
The move comes as AI agents are increasingly being developed to do more than simply answer questions. Instead of only generating text or providing information, these systems are being designed to interact with other services and perform actions based on user instructions.
Google's Home MCP rollout is currently focused on early adopters. The company said access will begin rolling out immediately and continue over the coming weeks to subscribers of Google Home Premium Advanced in the United States.
The subscription costs $20 per month and includes additional smart home features such as longer event-based video history, descriptive notifications, detailed alerts, video-history search tools and daily summaries.
Google has not said whether Home MCP will eventually become available to users on other Google Home subscription plans or in additional countries. The company also has not provided a timeline for a wider rollout.
During the early-access period, Google is asking users to provide feedback through its Smart Home for Developers Community. That feedback could help the company identify problems, improve the setup process and determine how the technology should develop before a broader release.
For now, the feature represents another step toward AI becoming an active interface for smart homes. Rather than controlling each connected device separately, users can increasingly rely on an AI agent to understand their instructions and interact with multiple parts of their home through a single conversation.
