MikesMagicMirror

A MagicMirror (v2.34.0) setup running on a Raspberry Pi with Ring camera integration, AI-powered object detection, and a two-page touchscreen dashboard.

Features

• Ring Camera Snapshots — Live snapshots from Ring doorbells and cameras, refreshed periodically and on motion events
• AI Object Detection — TFLite-based detection filters motion alerts to only show when a person, vehicle, or animal is in frame (SSD MobileNet v1 COCO)
• Two-Page Dashboard — Swipe between a main dashboard (clock, cameras, weather, indoor climate, calendar) and an info page (forecast, news, system stats)
• Touch Control — Tap to switch pages
• Remote Control — Web-based remote management with monitor on/off

Modules

Module	Description
MMM-RingSnapshot	Ring camera snapshots with AI person/vehicle/animal detection
MMM-DHT-Sensor	Indoor temperature and humidity from a DHT sensor
MMM-pages	Page/slide navigation for grouping modules
MMM-page-indicator	Visual indicator showing current page
MMM-Touch	Touch gesture support (tap to change pages)
MMM-Remote-Control	Web-based remote control interface
MMM-SystemStats	CPU temp, RAM, disk, and uptime display
MMM-PowerButtons	On-screen power/reboot/monitor controls

Setup

1. Install MagicMirror

Follow the official installation guide, or clone this repo:

git clone git@github.com:mbrande/MikesMagicMirror.git ~/MagicMirror
cd ~/MagicMirror
npm install

2. Configure

Copy the example config and fill in your values:

cp config/config.js.example config/config.js

You'll need to set:

• apiKey for MMM-Remote-Control
• Google Calendar private iCal URL
• Weather coordinates (lat/lon)

3. Ring Camera Setup

cd modules/MMM-RingSnapshot
npm install

Create a Ring refresh token (instructions) and save it:

cp ring-token.json.example ring-token.json
# Edit ring-token.json with your refresh token

4. AI Detection Setup

Run the one-time setup script to install the TFLite model and Python dependencies:

cd modules/MMM-RingSnapshot
bash setup_detection.sh

This creates a Python venv, installs ai-edge-litert, pillow, and numpy, and downloads the quantized SSD MobileNet v1 COCO model (~4MB).

Test it manually:

venv/bin/python3 detect_person.py <image.jpg> 0.5

5. Run

# With PM2 (recommended):
pm2 start npm --name magicmirror -- run start
pm2 save

# Or directly:
npm run start

AI Detection Config

These options go in the MMM-RingSnapshot config block in config/config.js:

Option	Default	Description
personDetection	true	Alert on person detection
personConfidence	0.5	Person confidence threshold (0-1)
vehicleDetection	true	Alert on vehicle detection (car, truck, bus, motorcycle, bicycle)
vehicleConfidence	0.5	Vehicle confidence threshold (0-1)
animalDetection	true	Alert on animal detection (cat, dog, bird, etc.)
animalConfidence	0.5	Animal confidence threshold (0-1)

Detection is fail-open: if the model errors or times out, motion alerts pass through normally. Detection only runs once at the start of a motion event, not during the 5-second refresh cycle.

Hardware

• Raspberry Pi 5
• Ring Doorbell / Camera
• DHT temperature/humidity sensor
• HDMI touchscreen display

License

MagicMirror is licensed under MIT. Based on MagicMirror by MagicMirrorOrg.