A tech enthusiast has successfully integrated local large language models (LLMs) with Home Assistant, a popular smart home platform, using the Home Assistant Model Context Protocol (MCP) server. This integration has significantly improved the management of smart devices, enabling voice commands to control various functions. The user experimented with different LLMs, ultimately finding that the Qwen 3.5 (9B) model provided the most reliable results for automation tasks. The unofficial HA-MCP Server repository was chosen for its ease of setup and compatibility with Home Assistant, despite initial hesitations about using community-supported tools. The integration allows for detailed control over smart home settings, automating tasks based on user queries, and managing devices effectively. However, the user has disabled certain high-access tools to prevent potential mishaps that could disrupt the smart home system. Overall, this project showcases the potential of combining local LLMs with smart home technology for enhanced user experience.