🥗 AI Nutritional Health Assistant — Personalized Guidance for Indian Diets

A conversational AI nutrition assistant specializing in Indian diets. Built with FastAPI, LangGraph, and a RAG (Retrieval-Augmented Generation) pipeline powered by FAISS vector search. Delivers context-aware nutrition guidance across multiple Indian regional cuisines with a modern Next.js 16 + TypeScript frontend and PostgreSQL database backend.

---

✨ Key Features

• 🍛 Regional Indian Cuisine Support — North, South, East, and West Indian cuisines
• 🤖 AI-Powered Recommendations — Personalized meal suggestions using RAG + LLM
• 💬 Conversational Interface — Natural language queries like "What should I eat for lunch in South India under 500 kcal?"
• 📊 Comprehensive Nutrition Data — Calories, macros (protein, carbs, fat), and micronutrients
• 🎯 Goal-Based Planning — Weight loss, maintenance, or muscle gain
• 🌱 Dietary Preferences — Vegan, vegetarian, non-vegetarian options
• 🏥 Health Condition Awareness — Allergy tracking and medical condition considerations
• 📱 Modern Web Interface — Responsive Next.js frontend with TypeScript
• 🐳 Docker Support — Easy deployment with Docker Compose
• 🔒 User Authentication — Secure login and profile management

---

🏗️ Architecture

Tech Stack

Backend

• FastAPI — High-performance async Python web framework
• LangGraph — Advanced workflow orchestration for AI agents
• LangChain — LLM application framework
• FAISS — Vector similarity search for RAG pipeline
• Sentence Transformers — Text embeddings (HuggingFace)
• PostgreSQL — User data and profile storage
• asyncpg — Async PostgreSQL adapter
• Pydantic — Data validation and settings management

Frontend

• Next.js 16 — React framework with App Router
• React 18 — Modern UI library
• TypeScript 5.0+ — Type-safe JavaScript

Infrastructure

• Docker & Docker Compose — Containerization with multi-container orchestration
• PostgreSQL 17 — Containerized database service
• Uvicorn — ASGI server for FastAPI
• NVIDIA API Endpoints — LLM inference (optional)

System Design

┌─────────────┐          ┌──────────────┐          ┌─────────────┐
│   Next.js   │  HTTP    │   FastAPI    │  Vector  │    FAISS    │
│  Frontend   │─────────▶│   Backend    │  Search  │   Index     │
│             │          │              │─────────▶│             │
└─────────────┘          └──────────────┘          └─────────────┘
                                │
                                │ SQL
                                ▼
                         ┌──────────────┐
                         │  PostgreSQL  │
                         │   Database   │
                         └──────────────┘

RAG Pipeline Flow:

1. User query → Embedding generation
2. FAISS vector search → Retrieve relevant food data
3. Context + Query → LLM prompt augmentation
4. LLM generates personalized nutrition advice
5. Response returned to user

---

📁 Project Structure

.
├── app/                        # FastAPI Backend
│   ├── routers/                # API route handlers
│   ├── services/               # Business logic layer
│   ├── models.py               # SQLAlchemy database models
│   ├── database.py             # Database configuration
│   └── main.py                 # FastAPI application entry point
│
├── frontend/                   # Next.js Frontend
│   ├── src/
│   │   ├── app/                # Next.js App Router pages
│   │   ├── components/         # Reusable React components
│   │   ├── lib/                # API client & utilities
│   │   └── styles/             # Global styles
│   └── package.json            # Node.js dependencies
│
├── faiss_RAG.py                # RAG pipeline implementation
├── food_dataset.csv            # Curated Indian food nutrition data
├── food_dataset.json           # JSON format food data
├── Food_dataset_Anuvaad.xlsx   # Regional dataset
├── requirements.txt            # Python dependencies
├── dockerfile                  # Docker image definition
├── docker-compose.yml          # Multi-container orchestration
├── start-servers.bat           # Quick start script (Windows)
├── .gitignore                  # Git ignore rules
└── README.md                   # This file

---

🚀 Getting Started

Prerequisites

• Python 3.9+ (3.10 or 3.11 recommended)
• Node.js 22+ and npm
• PostgreSQL 17 (or Docker for containerized deployment)
• Git
• .env file with database and API credentials (see Configuration section)

For Docker deployment: Only need Docker and Docker Compose

Quick Start with Docker

# Clone the repository
git clone https://github.com/theankitdash/AI-Nutritional-Health-Assistant-Personalized-Guidance-for-Indian-Diets.git
cd AI-Nutritional-Health-Assistant-Personalized-Guidance-for-Indian-Diets

# Create .env file with database credentials
cp .env.example .env  # (or create manually with required variables)

# Start all services with Docker Compose
docker-compose up --build

Services will be available at:

• Frontend: http://localhost:3000
• Backend API: http://localhost:8000
• API Docs: http://localhost:8000/docs
• Database: localhost:5432

Manual Setup (Without Docker)

Prerequisites

• Python 3.9+
• PostgreSQL 17 (must be running before starting backend)
• Node.js 22+

1. Backend Setup

# Navigate to project root
cd AI-Nutritional-Health-Assistant-Personalized-Guidance-for-Indian-Diets

# Install Python dependencies
pip install -r app/requirements.txt

# Create and configure .env file (see Configuration section below)

# Start FastAPI server
uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

✅ Backend running at: http://localhost:8000
📚 API Documentation: http://localhost:8000/docs

2. Frontend Setup (in new terminal)

# Navigate to frontend directory
cd frontend

# Install Node.js dependencies
npm install

# Start Next.js development server
npm run dev

✅ Frontend running at: http://localhost:3000

3. Database Setup

PostgreSQL must be running before starting the backend:

# On Windows (if PostgreSQL installed locally):
# PostgreSQL service should auto-start or start from Services

# On macOS/Linux:
brew services start postgresql
# or
sudo systemctl start postgresql

# Create database (optional, can be auto-created):
createdb nutrify_db

---

⚙️ Configuration

Environment Variables

Create a .env file in the project root with the following variables:

# Database Configuration
DB_NAME=nutrify_db
DB_USER=postgres
DB_PASSWORD=your_secure_password
DB_HOST=localhost          # Use 'postgres-db' for Docker
DB_PORT=5432

# API Configuration
API_BASE_URL=http://localhost:8000
NEXT_PUBLIC_API_URL=http://localhost:8000  # For Docker: http://fastapi:8000

# LLM Configuration (Optional - for NVIDIA API integration)
NVIDIA_API_KEY=your_nvidia_api_key

Docker Environment

When using Docker Compose, update the .env file to use Docker service names:

# For Docker services, use service names instead of localhost
DB_HOST=postgres-db
NEXT_PUBLIC_API_URL=http://fastapi:8000

🌐 Using the Application

1. Access the Application

Navigate to http://localhost:3000 in your browser

2. Register/Login

• Create a new account with email and password
• Or login with existing credentials
• Password is securely hashed with bcrypt

3. Complete Your Profile

Set up three key sections after logging in:

Personal Details

• Age/Date of Birth, gender, height, weight
• Waist circumference
• Activity level (sedentary, moderate, active)

Dietary Preferences

• Diet type (vegan, vegetarian, non-vegetarian)
• Regional cuisines (North, South, East, West Indian)
• Preferred meal types and ingredients
• Snack and sweet preferences
• Spice tolerance
• Caffeine intake
• Hydration level
• Meal frequency and eating out frequency

Health Conditions

• Food allergies and restrictions
• Medical conditions (diabetes, hypertension, etc.)
• Sleep quality and duration
• Supplement usage
• Fitness goals (weight loss, maintenance, muscle gain)

4. Start Chatting!

Ask the AI assistant questions like:

• "What should I eat for lunch in South India under 500 kcal?"
• "Give me a high-protein vegetarian dinner option"
• "I need a meal plan for weight loss with North Indian food"
• "What's a healthy breakfast option with less than 300 calories?"
• "Suggest meals for someone with dairy allergies"
• "I'm diabetic, what are safe Indian meal options?"

The assistant will provide personalized recommendations based on:

• Your dietary preferences
• Health conditions and allergies
• Fitness goals
• Regional cuisine preferences
• Nutritional requirements

---

📊 Dataset Information

The application uses curated Indian food nutrition datasets:

Data Files:

• food_dataset.csv — Primary nutrition database with Indian foods
• food_dataset.json — JSON format of food data
• Food_dataset_Anuvaad.xlsx — Extended regional cuisine data with translations

Data Attributes:

• Food name and aliases
• Calories and macronutrients (protein, carbohydrates, fat)
• Micronutrients (vitamins, minerals)
• Regional origin and cuisine type
• Common preparation methods

Data Sources:

• USDA FoodData Central
• Indian Food Composition Tables (IFCT)
• Regional Indian nutrition studies

FAISS Indexing:

• Food data is embedded using Sentence Transformers
• FAISS index stored in app/food_dataset/
• Enables fast semantic search for food recommendations

---

⚠️ Important Notes & Limitations

Warning

Medical Disclaimer: This application is for informational and educational purposes only and is NOT intended as medical or clinical nutrition advice. Always consult licensed healthcare professionals (doctors, registered dietitians) for:

• Medical nutrition therapy
• Chronic disease management
• Severe allergies or food sensitivities
• Personalized medical treatment plans

Known Limitations:

• Data Accuracy: Nutrition data accuracy depends on dataset quality; preparation methods and ingredient sourcing affect values
• Regional Coverage: Currently focused on major Indian regions (North, South, East, West); some cuisines may have limited data
• Allergies: Always independently verify ingredients if you have severe allergies or food sensitivities
• Portion Sizes: Recommendations are approximate; actual portions depend on individual needs and cooking methods
• Individual Variation: Nutritional needs vary based on metabolism, health conditions, and medications
• LLM Limitations: AI recommendations may occasionally be inaccurate; always verify with nutritional references
• GPU Acceleration: Embeddings are faster with GPU support, but CPU execution is supported for moderate usage

Best Practices:

• Use this tool as a starting point for nutrition planning
• Cross-reference recommendations with official nutrition databases
• Keep your profile information updated for better recommendations
• Consult healthcare professionals for medical conditions
• Report any inaccurate nutritional data to help improve the system

---

🤝 Contributing

We welcome contributions! Here's how you can help:

Ways to Contribute

1. Dataset Improvements

   • Add more regional Indian foods
   • Improve nutrition data accuracy
   • Add recipe information

2. Code Contributions

   • Bug fixes
   • New features (see Roadmap)
   • Performance optimizations
   • Test coverage

3. Documentation

   • Fix typos or improve clarity
   • Add examples
   • Translate to other languages

Contribution Process

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes with clear, descriptive commits
4. Add tests if applicable
5. Update documentation
6. Submit a Pull Request

Please follow coding standards and include tests for new features.

---

Dataset Licenses

• USDA FoodData Central: Public Domain
• Indian Food Composition Tables: Check specific source licenses

---

Made with ❤️ for healthier Indian diets