Plant Manager

Plant Manager is an AI-assisted houseplant monitoring backend. ESP32 sensors publish readings over MQTT, the backend stores time-series data in SQLite, applies threshold checks, and sends alerts/digests with optional LLM diagnosis.

Requirements

• Python 3.12+
• MQTT broker (e.g. Mosquitto)

Quick Start

1. Install dependencies:
   • pip install -r requirements.txt
   • or uv sync
2. Copy env template and configure:
   • copy .env.example .env (Windows)
3. Run backend:
   • python -m backend.main (preferred)
   • or python main.py
4. Run dashboard (in a separate terminal):
   • python -m dashboard.app
   • open http://localhost:5000

Configuration

Primary settings are in .env:

• LLM_PROVIDER (anthropic, openai, ollama)
• LLM_MODEL (optional; provider default used when empty)
• LLM_API_KEY (required for anthropic and openai)
• OLLAMA_HOST (used for ollama)
• MQTT_HOST, MQTT_PORT
• DB_PATH, PLANT_PROFILES_PATH
• SLACK_WEBHOOK_URL
• LOG_LEVEL
• DAILY_DIGEST_HOUR (0-23, UTC)
• ALERT_COOLDOWN_MINUTES (>=1)

MQTT Topic and Payload

• Topic format: plants/{location}/{plant_name}/sensors
• JSON payload format: {"moisture": 45, "temperature": 23.5, "humidity": 60, "light": 1200, "battery": 3.8}

Local Simulation

Use the simulator to test without hardware:

• python -m backend.simulator --once
• python -m backend.simulator --scenario dry-out
• python -m backend.simulator --test-notification

Dashboard Features

• Device tabs based on MQTT data source (location/plant_name)
• Current health status from profile threshold checks
• History charts for moisture, temperature, humidity, light, and battery
• Per-device plant profile assignment with automatic status range updates

Full Rebuild Guide

For complete end-to-end setup (Python env, Mosquitto, ESPHome flashing/calibration, backend/dashboard, and reboot automation), see:

• SETUP_GUIDE.md