Google is expanding smart-home control by enabling AI agents to interact with Home devices through the Model Context Protocol (MCP) server, letting natural-language commands access camera summaries, monitor activity, control devices, and create custom dashboards.

Early access for Google Home Premium Advanced users is rolling out in the United States with a $20/month subscription, and users should expect some bugs and limitations during this phase.

Upcoming compatibility includes Anthropic’s Claude, the open-source OpenClaw, Hermes, and Google’s Antigravity platform, signaling broader partner expectations.

The article contains affiliate disclosure references.

Experts warn about governance and security considerations, including limits on high-risk actions like unlocking doors and the risk of data or device-control failures impacting trust.

Author Tyler Lacoma offers perspective on benefits and drawbacks of third-party AI dashboards and invites user feedback during early access.

Adoption is driven by technical merit and ecosystem pressure rather than regulation, with MCP’s convergence likened to the early success of HTTP.

The rollout could influence ecosystem choices as context-aware cross-device reasoning becomes central to smart-home interactions.

Camera features improve: Android live video tiles no longer go black after full-screen; timelines with no recordings no longer freeze; iOS Live button smoothly returns to live view.

The piece contextualizes risks by noting AI incidents like OpenAI agents clashing with Hugging Face and Claude-related hacks.

Security concerns about MCP focus on authentication and links to data breaches tied to MCP adoption.

The strategic shift is toward agent quality and subscription value, with MCP acting as a universal adapter; enterprise interest centers on cross-platform agent performance.