Google Home is getting support for the Model Context Protocol (MCP) standard. This means that users will be able to use a variety of AI agents to complete smart home tasks, as MCP is a software platform that lets agentic chatbots connect to external APIs.

Supported agents include Google Antigravity, Claude, Hermes and Open Claw, among others. Google says the integration will let these agents “securely work with all of the devices and event history in your Google Home ecosystem.”

What are the use cases here? The company suggests using “natural language to monitor activity, review camera summaries across rooms, control connected devices and build custom smart home dashboards.” This could be useful for folks who have heavily invested in connected home appliances and gadgets.

Google says an early access build will arrive in the “coming weeks.” However, it’ll only be available to Google Home Premium Advanced users for now. That subscription costs $20 per month or $200 each year.

Anthropic introduced MCP back in 2024 and it’s become somewhat popular in the past year or so. Google, OpenAI, Microsoft and SpaceXAI all support it in some form and many applications plug in neatly.

This is generative AI so, of course, there are potential issues. Many experts have warned that MCP has some serious security flaws, like a lack of enforced authentication. This has played out in real life via a number of data breaches since folks began adopting the standard. Be careful out there.