domestic-motion

Exposes real-time motion detection events from domestic IoT sensors through the Model Context Protocol, allowing LLM agents to subscribe to and react to motion triggers in home environments. Implements MCP resource subscription patterns to stream sensor state changes with low-latency event delivery, enabling agents to build context-aware automation workflows based on physical motion events.

Aggregates motion events from multiple sensors across defined room or zone boundaries, providing agents with a unified view of occupancy and movement patterns at the room level rather than individual sensor level. Implements spatial grouping logic that correlates sensor readings to logical home zones, reducing noise and enabling higher-level reasoning about which areas are occupied.

Maintains a time-windowed history of motion events and exposes pattern analysis capabilities, allowing agents to query historical motion data and detect occupancy patterns (e.g., 'motion in kitchen between 7-9am daily'). Implements event buffering with configurable retention windows and provides statistical summaries of motion frequency, duration, and temporal clustering.

Exposes motion sensor metadata (location, sensitivity, battery status, last-seen timestamp) and allows agents to query or update sensor configurations through MCP tools. Implements a configuration schema that maps sensor IDs to physical locations, sensor types, and operational parameters, enabling agents to understand sensor capabilities and health.

Provides MCP tools for agents to define and trigger automations based on motion events, such as turning on lights, adjusting thermostats, or sending notifications. Implements a rule-action pattern where agents can register motion-triggered rules and the server executes corresponding actions, with support for conditional logic (e.g., 'turn on lights only if it's dark').