Google Home Opens Its Ecosystem to Third-Party AI Agents
Google is launching early access to its Home MCP server on September 16, 2026, allowing third-party AI agents to manage devices and analyze camera history for US-based Advanced tier users. The feature, which requires a $20 monthly subscription and Google Cloud project setup, enables advanced automation while maintaining strict safety guards against sensitive actions.
According to GN technics/smarthome (en-US), the rollout is officially scheduled for September 16, 2026, and requires an active Google Home Premium Advanced subscription at $20 per month, with setup involving specific OAuth configurations in a Google Cloud project.
Source: unite.aiAccording to GN technics/smarthome (en-US), the update highlights specific high-level capabilities like cross-camera activity summaries and automated voice notifications, while explicitly noting that the protocol blocks sensitive actions such as unlocking doors. The report also lists specific compatible third-party agents, including Antigravity, Claude, and OpenClaw, that can now connect to the system.
Source: 9to5google.comAccording to GN technics/smarthome (en-US), the setup process requires users to create a project in Google Cloud and link their specific AI agent—such as Claude or ChatGPT—before granting account permissions, with detailed instructions to be published in the Google Home Developer Center.
Source: Межа. Новини України.According to GN technics/smarthome (en-US), the integration specifically utilizes the Model Context Protocol to enable actions like inspecting home structures and analyzing historical events, with named compatibility for agents such as Claude and OpenClaw.
Source: XDAGoogle is allowing external AI systems to control smart home devices, but users must pay for a premium tier and accept that Gemini remains the default voice assistant.
Source: XDA Developers
