AI Document Processing System

A web application that automates processing and responding to Pre-Qualification Questionnaires (PQQs) and similar business documents. The system uses AI to extract form fields, match with previous responses, and generate completed documents.

Technology Stack

Backend

• Flask: Python web framework for API endpoints
• MCPs (Multimodal Collaborative Processors): Document processing engine for advanced document analysis
• Vision Models: Document visual analysis with configurable models (default: Gemini Flash)
• AI Document Processing: Intelligent field extraction and response generation
• Docx & Excel Processing: Native handling of Microsoft Office formats
• Session Management: Flask-Session for maintaining state during processing

Frontend

• React: Modern component-based UI framework
• Tailwind CSS: Utility-first CSS framework for styling
• JavaScript ES6+: Modern JavaScript for frontend logic
• Fetch API: For communication with the backend

AI Features

• Document field extraction with vision models
• Semantic matching with historical responses
• Contextual response generation
• Evidence document management
• Intelligent form completion

Project Structure

• backend/: Flask API server with document processing logic
• ppq-processing-system/: React frontend application
• example-rfp-responses-and-docs/: Sample documents for AI matching
• uploads/: Uploaded document storage
• output/: Generated output documents
• temp/: Temporary processing files
• flask_session/: Server-side session data

Setup Instructions

Prerequisites

• Python 3.9+ with pip
• Node.js 16+ with npm
• Git

Installation

1. Clone the repository:

   git clone https://github.com/matthewdcage/ai-document-processing.git
   cd ai-document-processing
   

2. Install backend dependencies:

   pip install -r requirements.txt
   

3. Install frontend dependencies:

   cd ppq-processing-system
   npm install
   cd ..
   

4. Create necessary directories:

   mkdir -p uploads output temp flask_session
   

Starting the Application

Run the provided script to start both backend and frontend servers:

./start_servers.sh

This will:

• Start the Flask backend server on http://localhost:5001
• Start the React development server on http://localhost:3000

Alternatively, start each component separately:

Backend:

cd backend
python api.py

Frontend:

cd ppq-processing-system
npm start

Usage

1. Navigate to http://localhost:3000 in your browser
2. Upload a document (PQQ, RFP, etc.)
3. Review the extracted fields and AI-suggested responses
4. Edit responses as needed
5. Submit the form to generate completed documents
6. Download or email the completed documents

Session Management

The system uses server-side sessions to maintain state during document processing. Files are stored in session-specific folders for security and organization.

Key Features

• Document field extraction using vision models
• AI matching of past responses from similar documents
• Document completion with user-edited values
• Evidence document management and attachment
• Email delivery of completed documents
• Session-based file management
• Timestamped output files

Contact

For more information, connect with me on LinkedIn.