Giving Nature a Voice through IoT & AI

Sprout is an intelligent plant monitoring system that bridges the gap between nature and technology. By combining real-time IoT sensor data with Google’s Gemini AI, Sprout transforms raw environmental metrics into a casual, friendly conversation with your houseplant.

---

🚀 Overview

Sprout isn't just a dashboard; it's a personality. The system monitors Moisture, Temperature, Humidity, and Light, feeding this data into a sophisticated 6-tier AI backend. Your plant doesn't just show you graphs—it tells you how it feels, gives you care tips, and provides a weekly "Botanist's Report" on its health.

---

✨ Key Features

• Live Vitals Dashboard: Real-time data visualization connecting physical sensors to a Flutter frontend via Bluetooth Classic.
• Casual Chat Interface: A "roommate-style" conversation engine that avoids robotic jargon.
• 6-Tier AI Fallback: Automated cascading logic using Gemini models to ensure 100% uptime.
• Smart Database Logging: Optimizes MongoDB storage by only logging significant environmental shifts, avoiding data bloat.
• Glassmorphism UI: A modern, clean aesthetic using frosted glass effects and organic colors.

---

🔌 Hardware Architecture

Sprout is built on a robust, logic-level protected microcontroller architecture.

Components & Signal Routing

• Microcontroller: Arduino Uno Rev3
• Wireless Communication: HC-05 Bluetooth Module (Classic SPP)
• Analog Sensors (Continuous Data via ADC):
  • Resistive Soil Moisture Sensor (Pin A0)
  • Photoresistor / LDR (Pin A1)
• Digital Sensors (Binary Data Streams):
  • DHT11 (Pin 2) for ambient Temperature & Humidity.

🛠️ Engineering Note — Logic Level Protection: > The architecture features a custom 1kΩ/4kΩ Voltage Divider to safely step down the Arduino's 5V TX signal to the HC-05's 3.3V RX logic level, preventing long-term hardware degradation. Additionally, a 4kΩ pull-down resistor is utilized for precise light level calibration on the LDR.

---

💻 Software & Backend

Component	Technology	Description
Backend Framework	Python / FastAPI	High-performance, asynchronous REST API.
AI Engine	Google GenAI SDK	Powers the conversational engine using Gemini.
Database	MongoDB Atlas	Motor Async driver for historical health logs.
Frontend App	Flutter / Dart	Cross-platform mobile UI with FL Chart for biological trends.
Connectivity	Bluetooth Serial	Standard SPP communication for local telemetry.

---

🧠 AI Strategy: The 6-Tier Fallback

To ensure Sprout never "loses its mind," the backend implements a specific priority sequence for AI requests. If a model hits a rate limit or is overloaded, the system automatically attempts the next tier:

1. Gemini 2.5 Flash Lite (Primary / Fast)
2. Gemini 2.5 Flash (Stable)
3. Gemini 2.0 Flash (Next-Gen Stable)
4. Gemini 2.0 Flash Lite (High Volume / Unlimited TPM)
5. Gemini 2.0 Flash Exp (Experimental Performance)
6. Gemini 2.5 Pro (Deep Reasoning / Complex Diagnostics)

---

📁 Project Structure

sprout_project/
├── sprout_backend/         # Python FastAPI Backend
│   ├── main.py             # API Endpoints & Routing
│   ├── ai_service.py       # Gemini Integration & Fallback Logic
│   ├── database.py         # Smart Logging & History Retrieval
│   └── .env                # API Keys (Keep Secret!)
├── sprout_app/             # Flutter Mobile Application
│   ├── lib/
│   │   ├── screens/        # Dashboard & Chat Screens
│   │   ├── services/       # Bluetooth Classic & API Integration
│   │   └── theme/          # Sprout Organic Design System
└── sprout_hardware/        # Arduino Firmware
    └── sprout_ble.ino      # Sensor Polling & JSON Serial Output

⚙️ Installation & Setup

1. Backend Setup

Navigate to the backend directory and install the required dependencies:

cd sprout_backend
pip install fastapi uvicorn google-genai python-dotenv motor

Create a .env file and add your configuration:

GEMINI_KEY=your_google_api_key_here
MONGO_URI=your_mongodb_connection_string

Run the server:

uvicorn main:app --reload

2. Mobile App Setup

Ensure you have the Flutter SDK installed, then navigate to the app directo

cd sprout_app
flutter pub get
flutter run

💬 Conversation Style (The "Vibe")

Sprout is designed to be your friend. It avoids "Assistant-speak" (like "How can I help you?"). Instead, it utilizes its system prompt to speak in casual, everyday English based on its current telemetry:

User: "How are you doing?"
Sprout (Moisture 10%): "Honestly? I'm pretty parched. A little water would be amazing right now."
Sprout (Healthy): "I'm having a great day. Just vibing in the sunlight. You look good today too!"

---

🧪 Future Roadmap

• Plant Vision: Disease detection using Gemini Multimodal Vision capabilities.
• Hardware Upgrade: Transitioning to Capacitive Soil Sensors for long-term corrosion resistance in wet soil.
• Auto-Watering: Integration with 5V submersible water pumps and relays for autonomous care.
• Social Sprout: Let multiple plants talk to each other over the local MQTT network.