Document Processing SaaS Platform

A full-stack intelligent document processing platform that transforms PDFs and images into structured data using AI/ML technologies.

Overview

This project demonstrates a production-ready SaaS platform that automatically processes documents using advanced AI/ML techniques. The system extracts text from PDFs and images, classifies documents by type, and identifies named entities - all through an intuitive web interface.

Architecture

┌─────────────────┐    HTTP/JSON     ┌──────────────────┐
│   Frontend      │ ◄──────────────► │   FastAPI        │
│   (HTML/JS)     │                  │   Backend        │
└─────────────────┘                  └──────────────────┘
                                              │
                                              ▼
                                    ┌──────────────────┐
                                    │  Processing      │
                                    │  Pipeline        │
                                    │                  │
                                    │ • PDF Parser     │
                                    │ • OCR Engine     │
                                    │ • ML Classifier  │
                                    │ • NER Extractor  │
                                    └──────────────────┘

Key Features

• Intelligent Text Extraction: Handles both text-based and scanned PDFs using OCR
• Document Classification: ML-powered categorization (invoices, contracts, letters, reports)
• Named Entity Recognition: Extracts people, organizations, dates, and monetary values
• Multi-language Support: Optimized for Italian text processing
• Responsive UI: Modern drag-and-drop interface with real-time progress
• Production Ready: Containerized with Docker for easy deployment

Technology Stack

Backend

• FastAPI - High-performance Python web framework
• spaCy - NLP library for entity recognition
• Tesseract OCR - Optical character recognition engine
• scikit-learn - Machine learning for document classification
• pdfplumber - PDF text extraction and parsing
• pdf2image - PDF to image conversion for OCR

Frontend

• Vanilla JavaScript - Modern ES6+ features
• HTML5/CSS3 - Semantic markup and responsive design
• Fetch API - Asynchronous HTTP requests

DevOps & Deployment

• Docker - Containerization for consistent deployments
• Uvicorn - ASGI server for production
• CORS - Cross-origin resource sharing configuration

Prerequisites

• Python 3.9+
• Docker (optional)
• Tesseract OCR
• Poppler utilities