Google announced that it’s giving Google Home access to any agent that supports calling Model Context Protocol (MCP). This includes ChatGPT, Claude, Hermes, and OpenClaw. The agents can control devices, analyze summaries, and build custom dashboards. It also lets an AI agent use a Google Home speaker as its “voice.”
A bigger neighborhood for Google Home
The abilities extend to all devices in the Google Home ecosystem, including Nest devices, Matter devices, and anything labeled “Works with Google Home.”
The announcement was made in a blog post by Taylor Lehman, product manager atr Google Home & Nest. Lehman said it is “just the next step in unlocking agentic smart homes.”
Right now though, access is only available to those in the US with its $20 per month Google Home Premium Advanced subscription. Google said it’s “rolling out,” which generally means about one to two weeks for it to extend to all eligible devices and users.
How to set up Google Home MCP
If you have a Google Home Premium Advanced subscription and want to get started with using your AI agent of choice, you’ll need to create a Google Cloud project.
Then you can configure it to use MCP by providing the MCP configuration details to your agent and asking it to set it up. The agent should then ask you to sign in and grant permissions.
Complete setup guides are available in the Google Home Developer Center.
Opens up homes to new security concerns
Giving an AI agent access to smart home controls doesn’t come without risks. Privacy and security issues abound when you’re talking about devices that watch and listen outside and inside homes, control lights and appliances, and run HVAC systems.
The post said that Home MCP does prohibit some sensitive actions, such as unlocking doors. “However, depending on your agent, connecting it to Home MCP can result in unexpected or even undesired behavior,” Lehman wrote.
