Google has begun rolling out early access to a Model Context Protocol (MCP) server for its Google Home ecosystem, according to TechCrunch and The Verge. The integration allows any AI agent that supports MCP — including Claude, Hermes, OpenClaw, ChatGPT, and Google Antigravity — to work with devices and event history in a user's Google Home setup. Taylor Lehman, group product manager at Google Home & Nest, described the capabilities in a blog post, including cross-camera analysis, reviewing device state history, sending audio messages over Home speakers, and building custom dashboards. The system supports Nest doorbells and thermostats and 'Works with Google Home' or Matter devices such as light bulbs. To set it up, users create a Google Cloud project, configure it to use Home MCP, and grant their chosen agent permissions. Google says Home MCP enforces rate limits and safety protections — for example, it will not allow an agent to unlock doors — but Lehman warned that 'connecting it to Home MCP can result in unexpected or even undesired behavior.' Access is initially limited to Google Home Premium Advanced subscribers in the US ($20/month or $200/year), rolling out over the coming weeks. Google would not comment on whether or when the feature would expand to other tiers or markets. MCP integration does not replace Gemini for Home, which remains Google's primary Home interface.
- Early access rolling out to Google Home Premium Advanced US subscribers ($20/month or $200/year)
- Supported agents include Claude, Hermes, OpenClaw, ChatGPT, and Google Antigravity
- Setup requires creating a Google Cloud project configured for Home MCP
- Safety limits included; agents cannot unlock doors, per Google
What it means for you
Google is letting outside AI assistants like Claude or ChatGPT read data from and control your smart home — cameras, thermostats, lights — instead of only Google's own assistant. In plain terms, you could ask an AI agent 'how many times did the front door open today?' or have it build a control dashboard for your house. It's an early, technical feature limited to Google's most expensive Home subscription in the US, so it's more experiment than everyday tool right now.
Try this
If you already pay for Google Home Premium Advanced and are comfortable with Google Cloud, follow Google's Home MCP setup guide to connect one agent and test a low-risk task like summarizing device activity — before giving it broader access.
Who should care
US smart-home hobbyists on Google's $20/month Premium Advanced tier who already use an AI agent and want to experiment with automating or querying their devices.
Skip this if
You don't use Google Home, aren't on the Premium Advanced tier, or aren't comfortable setting up a Google Cloud project — this isn't ready for casual users yet.
Sources: TechCrunch AI, The Verge AI — read the original