🧠 NeuroDoc – Intelligent Offline Document Analyzer

NeuroDoc is a Flask-based offline AI application designed to analyze documents (PDFs/images) using OCR and summarization models. It extracts text from uploaded documents, generates human-readable summaries, and allows users to download them—all without requiring internet access during inference.

• 📄 Upload PDF or image documents
• 🔍 Extract text using OCR (Tesseract)
• 🧠 Summarize text using a Transformer-based model (facebook/bart-large-cnn)
• 💾 Save and download summaries
• 🌐 Web-based interface built with Flask
• 📴 Runs offline after model download

NeuroDoc/
│
├── app.py                          # Main Flask backend
├── requirements.txt               # Python dependencies
├── Procfile                       # For Render deployment
├── render.yaml                    # Optional: Render config
│
├── templates/
│   └── index.html                 # Main frontend UI
│
├── static/                        # Static files (CSS/JS)
│
├── uploads/                       # Uploaded documents (auto-created)
├── summaries/                     # Generated summaries (auto-created)
│
├── neurodoc_core/                 # Core logic
│   ├── ocr_engine.py              # OCR using PyMuPDF + Tesseract
│   └── summarizer.py              # Summarization logic

Prerequisites:

• Python 3.8+
• Tesseract OCR installed on system

# Clone the repo
git clone https://github.com/BABIN-JOE/NeuroDoc.git
cd NeuroDoc

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

# Install Python dependencies
pip install -r requirements.txt

# Run the app
python app.py

Then open http://localhost:5000 in your browser.

You can download and cache the BART model beforehand to make the app work fully offline.

1. Go to https://render.com
2. Click "New Web Service"
3. Connect your GitHub repo
4. Set:
   • Build command: pip install -r requirements.txt
   • Start command: gunicorn app:app
5. Select Free Plan
6. Click Deploy

Make sure your requirements.txt and Procfile are present at the root.

• PDF
• JPG / PNG / BMP
• TIFF (if Tesseract supports it)

• Summaries and uploads are stored locally in /summaries and /uploads
• You can customize output format, compression, or add highlight detection easily
• Make sure Tesseract is installed and added to PATH

Babin Joe
🌐 Portfolio • GitHub • LinkedIn

This project is open-source and available under the MIT License.