🤖 Multi-LLM Architecture

Seamlessly switch between local language models (Llama3, Mistral, Gemma) directly from the mobile interface. No cloud dependencies—all inference happens on your machine.

🔍 Advanced Document Processing

Production-grade OCR pipeline using Tesseract for scanned PDFs and complex document layouts. Supports:

• Native text extraction from digital PDFs
• OCR processing for scanned documents and images
• DOCX and TXT file parsing
• Automatic text chunking with overlap for context preservation

🎯 Semantic Search Engine

FAISS-powered vector similarity search retrieves the most relevant document sections for each query, ensuring precise, context-aware answers with source attribution.

🔒 Privacy-Centric Design

Zero cloud dependencies for document processing and AI inference. Your documents never leave your local environment:

• Documents stored locally with PostgreSQL
• Vector embeddings computed on-device
• LLM inference through local Ollama server
• Complete data sovereignty

📱 Premium Mobile Interface

Modern React Native UI featuring:

• Real-time streaming AI responses with typing indicators
• Glassmorphism design language with smooth animations
• Intuitive document management and upload
• Conversation history with persistent storage
• Dynamic model selection and configuration

📊 Source Attribution

Complete transparency with document-level source tracking. Every AI response includes references to the specific documents used, maintaining academic rigor and trustworthiness.

📋 System Requirements

Backend Requirements

• Docker & Docker Compose (v2.0+) OR
• Python 3.9+ for manual setup
• PostgreSQL 13+ (included in Docker setup)
• Ollama installed and running locally
• Tesseract OCR 4.x

Frontend Requirements

• Node.js 18+ and npm
• Expo Go app (iOS/Android) for mobile testing
• Android Studio (for building standalone APK)

�️ Deployment Guide

Step 1: Ollama Configuration

Critical: Configure Ollama for network access to enable Docker communication.

# Pull your preferred language model
ollama pull llama3

# Enable network access (choose one method):

# Method 1: Via Ollama Settings UI
# Navigate to Settings → Enable "Expose Ollama to network"

# Method 2: Via environment variable
export OLLAMA_HOST=0.0.0.0
# Then restart the Ollama service

Available Models:

• llama3 - Meta's Llama 3 (Recommended)
• mistral - Mistral 7B
• gemma - Google's Gemma 2B/7B

Step 2: Backend Deployment

Option A: Docker (Recommended)

# Clone the repository
git clone https://github.com/Edge-Explorer/DOCUMIND-AI.git
cd DOCUMIND-AI

# Launch all services (backend + PostgreSQL)
docker-compose up --build

# Backend accessible at http://localhost:5000
# Database schema automatically initialized

The Docker setup includes:

• Flask API server
• PostgreSQL database with persistent volumes
• Automatic schema migrations
• Health check endpoints

Option B: Manual Setup

# Install Python dependencies
pip install -r requirements.txt

# Configure environment variables
cp .env.example .env
# Edit .env with your configuration:
# - DATABASE_URL=postgresql://user:password@localhost:5432/documind
# - SECRET_KEY=your-secret-key-here
# - OLLAMA_BASE_URL=http://localhost:11434

# Initialize database schema
python setup_db.py

# Start Flask development server
python app.py

Step 3: Frontend Setup

# Navigate to frontend directory
cd documind-frontend

# Install dependencies
npm install

# Configure API endpoint
# Edit src/api/config.ts and set:
# - API_BASE_URL to your local IP (e.g., http://192.168.1.100:5000)
# - For Docker setup, use your computer's local network IP

# Start Expo development server
npx expo start

# Scan QR code with Expo Go app on your mobile device

Finding Your Local IP:

• Windows: ipconfig (look for IPv4 Address)
• macOS/Linux: ifconfig or ip addr show
• Important: Use local network IP (192.168.x.x), NOT localhost

📁 Project Structure

DOCUMIND-AI/
├── app.py                 # Flask application entry point
├── models.py              # SQLAlchemy database models
├── database.py            # Database connection and session management
├── ollama_llm.py         # Ollama API integration and LLM wrapper
├── setup_db.py           # Database initialization script
├── docker-compose.yml    # Docker orchestration configuration
├── Dockerfile            # Backend container definition
├── requirements.txt      # Python dependencies
│
├── routes/               # API endpoint blueprints
│   ├── auth.py          # Authentication endpoints (register, login)
│   ├── documents.py     # Document upload and management
│   ├── chat.py          # Q&A conversation endpoints
│   └── models.py        # Model selection and management
│
├── ingest/              # Document processing pipeline
│   ├── processor.py     # Multi-format document parser
│   └── indexer.py       # FAISS vector store management
│
├── qa/                  # Question-answering logic
│   ├── retriever.py     # Semantic search and context retrieval
│   └── chain.py         # LangChain orchestration
│
├── migrations/          # Database migration scripts
│   └── alembic/         # Alembic migration versions
│
└── documind-frontend/   # React Native mobile application
    ├── src/
    │   ├── components/  # Reusable UI components
    │   ├── screens/     # Main app screens
    │   ├── api/         # API client configuration
    │   └── services/    # Business logic and state management
    ├── assets/          # Images, fonts, and icons
    └── app.json         # Expo configuration

🔧 Configuration

Environment Variables

Create a .env file in the root directory:

# Database Configuration
DATABASE_URL=postgresql://user:password@localhost:5432/documind

# Security
SECRET_KEY=your-secret-key-here
JWT_SECRET_KEY=your-jwt-secret-here

# Ollama Configuration
OLLAMA_BASE_URL=http://localhost:11434
DEFAULT_MODEL=llama3

# Application Settings
UPLOAD_FOLDER=./documents
VECTOR_STORE_PATH=./vector_stores

Docker Configuration

The docker-compose.yml file orchestrates two services:

services:
  db:
    image: postgres:15
    environment:
      POSTGRES_DB: documind
      POSTGRES_USER: documind
      POSTGRES_PASSWORD: documind123
    volumes:
      - postgres_data:/var/lib/postgresql/data
    ports:
      - "5432:5432"

  backend:
    build: .
    ports:
      - "5000:5000"
    environment:
      DATABASE_URL: postgresql://documind:documind123@db:5432/documind
      OLLAMA_BASE_URL: http://host.docker.internal:11434
    depends_on:
      - db

🎓 API Documentation

Authentication Endpoints

Register User

POST /api/auth/register
Content-Type: application/json

{
  "username": "john_doe",
  "email": "john@example.com",
  "password": "secure_password"
}

Login

POST /api/auth/login
Content-Type: application/json

{
  "email": "john@example.com",
  "password": "secure_password"
}

Response:
{
  "access_token": "eyJ0eXAiOiJKV1QiLCJhbGc...",
  "user": {
    "id": 1,
    "username": "john_doe",
    "email": "john@example.com"
  }
}

Document Management

Upload Document

POST /api/documents/upload
Authorization: Bearer <token>
Content-Type: multipart/form-data

file: <document.pdf>

List Documents

GET /api/documents
Authorization: Bearer <token>

Response:
{
  "documents": [
    {
      "id": 1,
      "filename": "research_paper.pdf",
      "upload_date": "2024-01-20T10:30:00Z",
      "file_type": "pdf",
      "page_count": 25
    }
  ]
}

Question Answering

Ask Question

POST /api/chat/ask
Authorization: Bearer <token>
Content-Type: application/json

{
  "question": "What are the main findings of the research?",
  "model": "llama3"
}

Response:
{
  "answer": "The research findings indicate...",
  "sources": [
    {
      "document": "research_paper.pdf",
      "page": 5,
      "relevance_score": 0.89
    }
  ]
}

🛠️ Development Workflow

Running Tests

# Backend tests
pytest tests/ -v

# Frontend tests
cd documind-frontend
npm test

Database Migrations

# Create new migration
alembic revision --autogenerate -m "Add new table"

# Apply migrations
alembic upgrade head

# Rollback migration
alembic downgrade -1

Building Production APK

cd documind-frontend

# Configure EAS Build
eas build:configure

# Build Android APK
eas build --platform android --profile production

# Build will be available in Expo dashboard

📊 Performance Optimization

Vector Store Optimization

The FAISS index uses HNSW (Hierarchical Navigable Small World) graphs for efficient similarity search:

• Chunk Size: 1000 characters with 200-character overlap
• Embedding Model: sentence-transformers/all-MiniLM-L6-v2
• Index Type: HNSW with M=16, efConstruction=200

Caching Strategy

• Vector embeddings: Cached in FAISS index files
• Document chunks: Stored in PostgreSQL for reuse
• LLM responses: Optional caching layer for repeated queries

🔒 Security Features

• JWT Authentication: Secure, stateless session management
• Password Hashing: bcrypt with configurable work factor
• SQL Injection Prevention: SQLAlchemy ORM parameterized queries
• File Upload Validation: Type checking and size limits
• CORS Configuration: Restricted to trusted origins
• Environment Isolation: Sensitive credentials in .env files

📜 Design Principles

• Privacy First: No external API calls for document processing or inference
• Modularity: Clean separation between ingestion, retrieval, and generation
• Scalability: Stateless API design for horizontal scaling
• Maintainability: Type hints, compr