What does MCP support in Google Home actually let an agent do?

An MCP-capable agent can query the home data Google Home already holds — device states and sensor readings — and issue control commands to linked devices. It is the same device list you see in the Home app, exposed to an external agent instead of only to Google's own assistant.

Do I need Gemini to use it?

No. The point of speaking MCP is that any MCP-capable client can use the integration; Gemini is one client among others, which is what makes the announcement different from a private vendor integration.

Is it available outside the United States?

Reporting says the rollout starts with Premium/Advanced subscribers in the US, with wider availability later. If your account is in another region, expect it to arrive after the US launch.

Google Home Adds MCP: AI Agents Can Now Control Your Smart Home

Google Home now supports the Model Context Protocol, so third-party AI agents can read home sensor data and control linked devices. Here is what shipped, who gets it first, and what it means if you already run an agent on your PC.

Google Home now supports the Model Context Protocol (MCP), which means third-party AI agents — not just Google’s own assistant — can read your home data and control linked devices. TechCrunch, The Verge and Engadget all describe the same shape: agents that speak MCP get a supported way into the smart home, and the rollout starts with Premium/Advanced subscribers in the US.

Key facts

Fact	Detail
Question answered	Google Home Adds MCP: AI Agents Can Now Control Your Smart Home
Short answer	Google Home now supports the Model Context Protocol, so third-party AI agents can read home sensor data and control linked devices.
Reading time	~2 min
Last updated	2026-09-17

What actually shipped

MCP support in Google Home, so an external agent can query the device and sensor data the Home app already holds.
Control, not just reading — the same integration is described as letting agents act on linked devices.
Staged rollout: Premium/Advanced subscribers in the US first.
The Verge frames it as Google letting “any AI agent” run the smart home, rather than reserving automation for Gemini.

Why MCP support is the interesting part

Before this, connecting an assistant to a smart home usually meant a private integration or a vendor-specific cloud API, and the assistant had to be the vendor’s own. MCP changes the unit of work: “an agent can control the house” becomes a tool any MCP-capable client can call — including an agent running on your own computer rather than in someone else’s cloud.

That is a meaningful shift for anyone who wants automation without handing the whole control plane to one vendor.

What this means if you run an agent on Windows

A tool that can read sensors and flip switches raises the permission question immediately. An agent that touches your home should say what it is about to do, ask first, and leave a record you can review — the same discipline you would want for any local action.

That is the model MOOGH uses for work on your PC: plan, ask for approval, then execute on your own machine. A smart home call is just one more permissioned tool. If you want that boundary on your own computer, download MOOGH for Windows.

FAQ

See the questions above — MCP support is about which agents may act, Gemini or otherwise, and the first wave is US Premium/Advanced accounts.

Related reading:

Google Home Now Speaks MCP — Third-Party AI Agents Can Read Your Home and Control Devices

Key facts

What actually shipped

Why MCP support is the interesting part

What this means if you run an agent on Windows

FAQ

Sources

Frequently asked questions

Blog