🌾 GraamSeva: AI-Powered Rural Empowerment 🇮🇳

Bridging the Digital Divide for 700M+ Citizens through Voice and AI.

GraamSeva is a state-of-the-art AI platform designed specifically for rural Indian farmers. It simplifies and democratizes access to government welfare schemes using advanced Large Language Models (LLM), Real-time Geolocation, and Voice-to-Voice Interaction.

---

✨ Key Features

🤖 Intelligent AI Chatbot (Gemini Powered)

The heart of GraamSeva is an LLM-driven Assistant that:

• Understands Context: "I have 10k for a tractor" → Identifies as a downpayment and suggests loans.
• Smart Redirection: Automatically navigates users to Mandis, Loans, or Scheme pages based on intent.
• Strict Localization: Responds in the user's preferred language (Hindi, Bhojpuri, Awadhi, Odia, Marathi, Maithili).

📍 Location-Aware Intelligence

• Nearby Discovery: Integrated with browser Geolocation to suggest the closest Mandis (markets) and Bank branches.
• Dynamic Data: Real-time price and distance estimation based on actual user coordinates.

🎤 Voice-First Interface

• Low Literacy Optimized: No typing required. Users speak in their native tongue.
• Web Speech API: Browser-native, high-performance transcription.

🌐 Regional Language Support

• Inclusive Access: Users can switch the entire interface to their preferred regional language.
• Full Interface Translation: Schemes, chatbot responses, and all translatable UI elements adapt automatically.
• Supported Languages: Hindi, Bhojpuri, Awadhi, Marathi, Maithili, Odia, and English.

📊 Operator Dashboard

• CSC Management: Real-time KPI monitoring, language distribution analytics, and application tracking for village operators.

---

🏗️ System Architecture

graph TD
    User((Farmer / User)) -->|Voice/Text| UI[React Frontend]
    UI -->|Query + Context| LLM[Google Gemini API]
    UI -->|Geolocation| Loc[Browser Geolocation API]
    
    subgraph "AI Content Engine"
    LLM --> Intent{Intent Router}
    Intent -->|Mandi Query| Mandi[Market Price Page]
    Intent -->|Loan Query| Loan[Loan Suggestion Page]
    Intent -->|Scheme Query| Schemes[Government Schemes Page]
    end
    
    subgraph "Local Storage & State"
    UI --> State[React Context/State]
    State --> Cache[LocalStorage Thread Caching]
    end
    
    Mandi --> DB[(Govt/Bank APIs)]
    Loan --> DB
    Schemes --> DB

Loading

---

🛠️ Technology Stack

Layer	Technology
Frontend	React 19, Vite, Tailwind CSS
Backend	Django 6.0 + Django REST Framework
AI / LLM	Google Gemini 2.5 Flash
Database	SQLite (Development), PostgreSQL (Production-ready)
Styling	Tailwind CSS, Responsive Grid
Icons	Material Design Icons

---

🚀 Quick Setup

1. Prerequisites

• Node.js (v18+)
• npm
• django

2. Installation

Frontend Setup

# Navigate to the frontend directory
cd GraamSevaFrontend

# Install dependencies
npm install

Backend Setup

# Navigate to the backend directory
cd Backend/GraamSevaBackend

# Create and activate virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install Python dependencies
pip install django djangorestframework

# Run migrations
python manage.py migrate

# Create superuser (optional, for admin panel)
python manage.py createsuperuser

3. Environment Configuration

Create a .env file in GraamSevaFrontend/ and add your Gemini API Key:

VITE_GEMINI_API_KEY="your_google_gemini_api_key"

4. Run Development Server

Start Frontend (React + Vite)

cd GraamSevaFrontend
npm run dev

Open http://localhost:5173 to explore.

Start Backend (Django REST API)

cd Backend/GraamSevaBackend
source venv/bin/activate
python manage.py runserver

Backend API available at http://localhost:8000
Django Admin panel: http://localhost:8000/admin

---

📂 Project Structure

rnicrosoftHackathon/
├── GraamSevaFrontend/          # React Frontend Application
│   ├── src/
│   │   ├── components/         # Reusable UI (AssistantBar, etc.)
│   │   ├── pages/              # 12+ Feature Pages (Mandi, Loan, Dashboard)
│   │   ├── services/           # LLM, Voice, API Client logic
│   │   ├── constants/          # App configuration & language maps
│   │   └── lib/                # i18n and Utility helpers
│   ├── .env                    # API Keys (Git ignored)
│   └── vite.config.js          # Build configuration
│
├── Backend/                    # Django REST API Backend
│   └── GraamSevaBackend/
│       ├── manage.py
│       ├── GraamSevaBackend/   # Project settings & configuration
│       │   ├── settings.py     # Django settings with REST framework config
│       │   ├── urls.py         # API route definitions
│       │   ├── wsgi.py         # Production WSGI server
│       │   └── asgi.py         # Async WSGI server
│       └── venv/               # Python virtual environment
│
└── README.md                   # This file

---

Backend API Integration

The Django REST API (Backend/GraamSevaBackend) provides core services:

Key API Endpoints (Planned/Available)

• Schemes: GET /api/schemes, /api/schemes/:id, /api/schemes/search
• Eligibility: POST /api/eligibility/check, GET /api/eligibility/criteria/:schemeId
• Applications: POST /api/applications, GET /api/applications/:id
• Intent Routing: POST /api/intent/route, /api/intent/classify
• Mandis: GET /api/mandis, /api/mandis/nearby
• Loans: GET /api/loans, POST /api/loans/compare

Frontend-Backend Communication

The React frontend communicates with the backend via apiClient.js and apiConfig.js:

• Base URL configured via VITE_API_BASE_URL environment variable
• Fallback to mock data when APIs are unavailable
• Error handling and retry logic built-in

---

• Untracked Secrets: .env is removed from Git history and excluded via .gitignore.
• API Safety: LLM responses are parsed strictly to prevent injection.

---

🏆 Acknowledgments

Built by Team Bada Pikachu (IIT Roorkee) for the AI Unlocked Hackathon.

"Serving the roots of India with the power of the cloud."