Google has announced that its Home platform will integrate with the Model Context Protocol (MCP), a standard designed to let AI agents connect securely to external APIs. This move enables users to deploy various agentic chatbots to manage their smart home ecosystems more intelligently.
According to Google, supported AI agents will include Google Antigravity, Claude, Hermes, and Open Claw. These agents will be able to interact with all devices within a user’s Google Home setup and access event history. The company highlighted several practical applications, such as monitoring activity through natural language queries, reviewing camera summaries across different rooms, controlling connected appliances, and constructing custom smart home dashboards.
The integration comes as MCP, introduced by Anthropic in 2024, gains traction across the tech industry. Major players including Google, OpenAI, Microsoft, and SpaceXAI have adopted the standard in some capacity, leading to widespread compatibility with numerous applications.
However, the shift toward agentic AI in home environments is not without risks. Security experts have flagged potential vulnerabilities in the MCP standard, specifically noting a lack of enforced authentication mechanisms. These concerns are not merely theoretical; there have been documented cases of data breaches linked to MCP adoption, prompting warnings for users to exercise caution.
Early access to this feature is planned for the coming weeks, but it will initially be restricted to Google Home Premium Advanced subscribers. The premium tier is priced at $20 per month or $200 annually.
$20 a month for an open API with known security flaws? I’m keeping my smart home dumb and safe.
Finally, a standard that actually works! Worth the $20 if my house can clean itself without me lifting a finger.