OmniDoc AI - Enterprise OCR Document Intelligence Platform

Developed by Hassan Khan
📧 Email: hassanaiengineer@gmail.com
📞 Phone: +923131569393 🔗 LinkedIn: https://www.linkedin.com/in/hassan-khan-4961b722b/
🔗 Upwork: https://www.upwork.com/freelancers/~016ca6a619d9683838

Overview

OmniDoc AI is an enterprise-grade OCR-first document intelligence platform that transforms messy scanned documents into clean, structured, AI-ready knowledge using Gemini 2.0 Flash and RAG (Retrieval-Augmented Generation).

Key Features

🔍 OCR-First Architecture

• Column-aware extraction with layout preservation
• Auto-detect digital vs scanned pages
• Pre-processing: deskew, denoise, contrast normalization
• Support for PDF, PNG, JPG, JPEG, TIFF, BMP
• Page-level confidence scoring

✨ AI-Powered Text Cleaning

• Simple Cleaner (Free): NLP-based preprocessing
  • Unicode normalization
  • OCR error correction
  • Whitespace & punctuation fixes
  • Broken word merging
• Advanced Cleaner (Gemini 2.0 Flash)
  • Context-aware corrections
  • Grammar & structure improvement
  • No hallucination mode

💬 Enterprise RAG

• Document Q&A with citations
• OCR-aware chunking
• Semantic search with relevance scoring
• Multi-document support
• Source references

🏢 Enterprise Features

• User-owned API keys (secure, no platform cost)
• Document isolation per project
• Multiple output formats (TXT, Markdown, JSON)
• Batch processing support

Tech Stack

Backend

• FastAPI (Python)
• PyMuPDF, OpenCV, Tesseract (OCR)
• Google Gemini 2.0 Flash (AI)
• ChromaDB (Vector Store)

Frontend

• Next.js 14
• React 18
• TypeScript
• Tailwind CSS
• Framer Motion

Getting Started

Prerequisites

• Python 3.10+
• Node.js 18+
• Tesseract OCR
• Poppler (for PDF processing)

Backend Setup

cd backend

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

# Install dependencies
pip install -r requirements.txt

# Run the server
python -m uvicorn app.main:app --reload --port 8000

Or use the provided script:

chmod +x run.sh
./run.sh

Frontend Setup

cd frontend

# Install dependencies
npm install

# Run development server
npm run dev

The frontend will be available at http://localhost:3000

System Dependencies

Ubuntu/Debian

sudo apt-get update
sudo apt-get install tesseract-ocr poppler-utils

macOS

brew install tesseract poppler

Windows

• Download Tesseract from: https://github.com/UB-Mannheim/tesseract/wiki
• Download Poppler from: https://github.com/oschwartz10612/poppler-windows/releases

API Endpoints

Documents

• POST /api/documents/upload - Upload a document
• POST /api/documents/upload-multiple - Upload multiple documents
• GET /api/documents/ - List all documents
• GET /api/documents/{id} - Get document details
• DELETE /api/documents/{id} - Delete a document
• GET /api/documents/{id}/download - Download document text

OCR

• POST /api/ocr/process/{id} - Process document with OCR
• POST /api/ocr/process-batch - Batch OCR processing
• POST /api/ocr/reprocess/{id} - Reprocess document
• GET /api/ocr/status/{id} - Get OCR status

Cleaning

• POST /api/cleaning/simple/{id} - Simple NLP cleaning
• POST /api/cleaning/advanced/{id} - AI-powered cleaning (requires Gemini API key)
• GET /api/cleaning/preview/{id} - Preview cleaning results

RAG

• POST /api/rag/index/{id} - Index document for RAG
• POST /api/rag/index-batch - Batch indexing
• POST /api/rag/query - Query documents
• DELETE /api/rag/index/{id} - Remove from index

Settings

• GET /api/settings/ - Get settings
• POST /api/settings/ - Update settings
• POST /api/settings/validate-gemini-key - Validate API key

Configuration

Environment Variables

Create a .env file in the backend directory:

# Optional: Set default API key (not recommended for production)
GEMINI_API_KEY=your_key_here

For the frontend, create .env.local:

NEXT_PUBLIC_API_URL=http://localhost:8000

Project Structure

omnidoc-ai/
├── backend/
│   ├── app/
│   │   ├── main.py           # FastAPI application
│   │   ├── models/
│   │   │   └── schemas.py    # Pydantic models
│   │   ├── services/
│   │   │   ├── ocr_service.py
│   │   │   ├── cleaning_service.py
│   │   │   ├── gemini_service.py
│   │   │   ├── rag_service.py
│   │   │   └── document_store.py
│   │   ├── routers/
│   │   │   ├── documents.py
│   │   │   ├── ocr.py
│   │   │   ├── cleaning.py
│   │   │   ├── rag.py
│   │   │   └── settings.py
│   │   └── utils/
│   ├── uploads/              # Uploaded files
│   ├── outputs/              # Processed outputs
│   ├── requirements.txt
│   └── run.sh
└── frontend/
    ├── src/
    │   ├── app/
    │   │   ├── page.tsx      # Main dashboard
    │   │   ├── layout.tsx
    │   │   └── globals.css
    │   ├── components/
    │   │   ├── Header.tsx
    │   │   ├── FileUploader.tsx
    │   │   ├── DocumentList.tsx
    │   │   ├── DocumentViewer.tsx
    │   │   ├── CleaningModal.tsx
    │   │   ├── RAGChat.tsx
    │   │   ├── SettingsModal.tsx
    │   │   ├── ProcessingStatus.tsx
    │   │   └── StatsCard.tsx
    │   └── lib/
    │       └── api.ts        # API client
    ├── package.json
    ├── tailwind.config.js
    └── next.config.js

Usage Guide

1. Upload Documents: Drag and drop or click to upload PDFs/images
2. OCR Processing: Documents are automatically processed after upload
3. Clean Text:
   • Click "Clean" on any document
   • Choose Simple (free) or Advanced (AI-powered)
   • Download the cleaned text
4. RAG Chat:
   • Configure your Gemini API key in Settings
   • Click "Index RAG" on documents
   • Go to RAG Chat tab
   • Ask questions about your documents

Security Notes

• API keys are stored locally in your browser
• Keys are sent directly to Google's servers
• No keys are stored on our backend
• Document data is stored locally during your session

License

MIT License - See LICENSE file for details

Support

For support or feature requests, please contact:

• 📧 Email: hassanaiengineer@gmail.com
• 📞 Phone: +923131569393

---

Note - If you want frontend folder send email to me :)