ha-analyze-patterns

Home Assistant pattern analysis reads historical snapshots, live context via MCP, and normalized inventories to identify unused devices, energy anomalies, and automation opportunities. It generates a Markdown summary and updates MEMORY.md with learned patterns.

You can find unused devices by analyzing Home Assistant entity patterns to identify devices that are unused, always-off, or intermittently active. This pattern analysis reads normalized inventories to highlight hardware removal or optimization opportunities.

The best way to detect energy usage anomalies is by analyzing Home Assistant historical data and entity patterns. This approach surfaces energy anomalies and proposes actionable usage insights for optimization and cost savings.

Home Assistant pattern analysis requires no external dependencies. It reads raw snapshots, live context via MCP, and normalized inventories directly from your environment to perform periodic health checks and optimization sessions across dashboards and automations.

Pattern analysis updates memory by writing pattern data to raw JSON files, creating a Markdown summary when non-trivial findings exist, and updating MEMORY.md with learned patterns. This ensures optimization insights and automation opportunities are tracked.