
LEGO-Anything shows promise, but benchmark results expose weak geometric self-assessment.
Google is rolling out early access to a Google Home MCP server that lets AI agents control connected devices, review camera summaries, and access smart home activity through natural language.

Google is rolling out early access to a Model Context Protocol server for the Google Home ecosystem. The update allows AI agents that support MCP — including Claude, Hermes, OpenClaw, ChatGPT, and Google Antigravity — to work with smart home devices and access event history. For users, the core shift is simple: smart home control can move from app taps and preset commands to natural-language requests.
The MCP connection is designed to let people review camera summaries, monitor smart home activity, control connected devices, and build custom smart home dashboards. Supported hardware includes devices in the Google Home ecosystem, Google Nest doorbells and thermostats, and Works with Google Home or Matter devices such as light bulbs. The practical takeaway: AI agents could become a more flexible interface for everyday home automation.
To connect an agent, users need to create a Google Cloud project and configure it for Home MCP. They then provide MCP configuration details to their chosen agent and ask it to set up the connection. The agent prompts the user to sign in and grant permissions, and Google says a setup guide will be available in the Google Home Developer Center.
Access is rolling out starting September 16, 2026, and continuing over the following weeks for U.S. subscribers to Google Home Premium Advanced. The article identifies this as the $20-per-month tier, with features such as longer event-based video history, descriptive notifications, detailed alerts, video history search tools, and daily summaries. Google did not comment on if or when the MCP will expand to other subscription tiers or markets.

LEGO-Anything shows promise, but benchmark results expose weak geometric self-assessment.

RRSI aims to help self-improving AI agents generalize beyond their test tasks while using fewer tokens.

A quick guide to the AI agents built for messaging, family logistics, travel, work, and everyday tasks.

Dots are always-on AI agents for Pro and Business Premium users in eligible markets.