SoilSong AI 🎵🌱

An AI-powered mobile app that tells the story of your soil.
Built for the 2025 Call for Code Hackathon – Theme: Decent Work and Economic Growth
Addresses: SDG 15 (Life on Land), SDG 8 (Decent Work and Economic Growth), and SDG 2 (Zero Hunger)

OLifie Logo

📱 App Overview

SoilSong AI transforms raw soil data into audio-based narratives that educate and empower farmers.
Users input basic soil metrics like pH and moisture—or upload a soil photo—and the app returns a spoken story and actionable advice to improve soil health and sustainable land use.

Designed for inclusivity and simplicity, especially in low-literacy, rural communities, the app speaks to farmers in a language they understand—literally and figuratively.

🌍 Sustainable Development Goals (SDGs) Addressed

SDG 15 – Life on Land
Encourages regenerative agriculture and protects soil ecosystems.
SDG 8 – Decent Work and Economic Growth
Enables better farming yields and promotes agritech-driven livelihoods.
SDG 2 – Zero Hunger
Supports improved soil fertility, contributing to higher food security.

🧠 Powered by IBM Granite (watsonx.ai)

SoilSong AI uses IBM Granite foundation models through watsonx.ai to turn structured soil inputs into rich, localized stories:

Converts numeric soil data (e.g., pH: 5.3, moisture: low) into human-readable messages like:
"Your soil is a bit sour and dry today. It could use some compost and water soon!"
Outputs text is tailored with prompt engineering to fit a narrative style.

🔜 Planned integrations (post-hackathon):

Multilingual output using Granite LLMs
Voice generation with Watson Text-to-Speech
Image analysis with Watson Visual Recognition for soil texture detection

🔧 Technologies Used

Frontend (React Native – Expo)

expo-image-picker – soil image selection
expo-av – audio playback
axios – API requests

Backend (To deploy: IBM Cloud Functions or Flask)

IBM watsonx.ai (Granite models) – story generation
IBM Watson Text-to-Speech – audio (future phase)
IBM Watson NLU – understanding soil context
IBM Cloud Hosting – scalable cloud deployment

📦 Installation

git clone 'repo'
cd soilsong-ai
npm install
npx expo start


Scan the QR code using your Expo Go mobile app.

🧪 Testing Instructions
Enter a sample soil pH value (e.g., 6.4) and moisture percentage (e.g., 18).

(Optional) Upload a soil image.

Tap “Generate Soil Story.”

The app returns:

A personalized narrative (text)

(In future) A spoken version of the story

📁 Folder Structure
bash
Copy
Edit
/screens
  ├── HomeScreen.js        # Input form + optional photo
  └── ResultScreen.js      # Displays generated story + audio
App.js                     # Navigation
🔗 IBM Services Setup (Backend)
To connect IBM Cloud to your backend, follow these steps:

Create a Free IBM Cloud account

Enable the following services:

watsonx.ai with Granite models (prompt-based story generation)

Watson Natural Language Understanding

Watson Text-to-Speech (for voice, optional)

Watson Visual Recognition (optional)

🚀 Project Status
✅ MVP Completed
⏳ Text-to-Speech, language translation, and image analysis planned for future updates

“When the soil speaks, the farmer listens.” –  SoilSong AI

vbnet
Copy
Edit