This is my production Home Assistant configuration powering a complete smart home setup with over 40 custom components, multi-protocol device integration, AI-powered video surveillance, and advanced automation scenarios. The repository is both a working configuration and a reference architecture for complex Home Assistant deployments.
- System Architecture
- Hardware Ecosystem
- 40+ Custom Components
- ESPHome Devices
- Frigate NVR: AI Video Surveillance
- Deployment: GitHub Actions Auto-Deploy
- Source Code
- FAQ
- What hardware is required?
- How is the configuration deployed?
- Can I use parts of this configuration?
- Why 40+ custom components instead of built-in integrations?
System Architecture
The Home Assistant configuration runs on a multi-service stack: HA Core as the central automation hub, InfluxDB/VictoriaMetrics for time-series metrics, Zigbee2MQTT for Zigbee devices, ESPHome for custom ESP32/ESP8266 controllers, Frigate NVR for AI-powered video surveillance, Node-RED for visual automation flows, and Glances for system monitoring.
The modular approach uses packages (ha_sensor.yaml, ha_switch.yaml, ha_light.yaml, telegram.yaml, xiaomi.yaml, yandex.yaml) to keep the configuration maintainable as the device count grows.
Hardware Ecosystem
The setup integrates devices across multiple protocols and ecosystems:
- Xiaomi ecosystem: Gateway 3, Yeelight LEDs, air purifiers, rice cookers, Dreame vacuum cleaners, BLE sensors
- Network: MikroTik routers, MiWiFi, OpenWRT devices
- Media: Samsung Tizen TVs, LG WebOS TVs, Bang & Olufsen Beoplay, Yandex Station speakers
- Appliances: Haier HON, Venta air washers, HyperHDR ambient lighting
- ESPHome devices: Adafruit ESP32-S3, M5Paper e-ink dashboard, custom RPM monitors, OLED displays
40+ Custom Components
The custom_components/ directory is extensive, covering everything from solar energy (Anker Solix), room occupancy detection, flight tracking (Flightradar24), Docker container management (Portainer), Proxmox VE monitoring, mesh networking (Meshtastic), to power consumption calculation (Powercalc). Each component was selected for production reliability and actively maintained.
ESPHome Devices
Custom ESPHome configurations include an Adafruit ESP32-S3 with sensor packages, an ESP32-S3 with 0.42″ OLED display for compact status indicators, the M5Paper e-ink display for a persistent dashboard, and an RPM monitor. The ESPHome directory includes reusable packages and custom components for display rendering.
Frigate NVR: AI Video Surveillance
Frigate provides object detection on camera feeds, identifying people, vehicles, and animals with real-time notifications. The configuration integrates with Home Assistant for automated responses: lights on when person detected, recording triggers, and Telegram alerts.
Deployment: GitHub Actions Auto-Deploy
A GitHub Actions workflow (.github/workflows/auto-deploy.yaml) enables continuous deployment, push to the repository and the configuration automatically deploys to the Home Assistant instance. This CI/CD approach to smart home management ensures version control, rollback capability, and collaborative editing.
Source Code
The configuration is available upon request. Contact me for access. As a Fractional CTO, I apply the same infrastructure-as-code principles to smart home automation that I use in enterprise environments.
For IoT and smart home consulting, 15 years of experience in system integration across 38 countries.
FAQ
What hardware is required?
Home Assistant runs on a Proxmox VM or dedicated device. A Zigbee coordinator (USB stick) is needed for Zigbee devices, and a Coral TPU is recommended for Frigate NVR object detection performance.
How is the configuration deployed?
Via GitHub Actions. Push changes to the repository and they automatically deploy to the HA instance. Manual deployment is also possible by pulling the repo directly on the HA host.
Can I use parts of this configuration?
Yes. The modular package structure means you can pick individual YAML files (e.g., the Telegram bot integration, Xiaomi configuration, or ESPHome device configs) and adapt them to your setup.
Why 40+ custom components instead of built-in integrations?
Many devices need functionality beyond core HA integrations. Custom components from HACS provide deeper control, like Dreame vacuum room-specific cleaning, MikroTik advanced routing metrics, or Proxmox VM management. Each was chosen for production stability.
