AI-Supported Invoice Processing Tool

This project automates invoice processing by fetching invoices from emails, extracting key details, and storing them in MongoDB. It uses Google Vision API to extract invoice information from PDF attachments.

Features

• Automatically fetches invoices from emails
• Extracts key invoice details (Invoice Number, Amount, Due Date, Sender)
• Processes PDF attachments using Google Vision API
• Stores extracted invoices in MongoDB
• Simple Web Dashboard for viewing and editing invoices

Technologies Used

• Python (Flask, pymongo, pdfplumber, IMAPClient)
• MongoDB Atlas (Database for invoices)
• Google Vision API (OCR for extracting text from PDFs)
• Flask (For the web dashboard)
• CSS (For styling)

Setup Guide

1. Prerequisites

   • Python 3.8+
   • MongoDB instance (local/Atlas)
   • Google Cloud Account (For Vision API)
   • Email account for IMAP Access

2. Installation

# Clone repository
git clone https://github.com/yourusername/invoice-processor.git
cd invoice-processor

# Install dependencies
pip install flask pymongo imapclient pdfplumber google-cloud-vision pillow python-dotenv

3. Congiguration Create .env file

# Email Configuration
IMAP_SERVER=imap.gmail.com
EMAIL_USER=your-email@gmail.com
EMAIL_PASS=your-email-password

# MongoDB Configuration (For MongoDB Atlas)
MONGO_URI=mongodb+srv://your-username:your-password@your-cluster.mongodb.net/?retryWrites=true&w=majority
MONGO_DB_NAME=ai-invoice
MONGO_COLLECTION_NAME=invoices

4. Google Vision Setup
   • Create service account credentials in Google cloud console
   • Download the JSON Key File
   • Save as google_vision_key.json in project root

Usage

Process Emails

python app.py

Access Dashboard

python dashboard.py

Visit http://localhost:5000 to:

• View all the invoices
• Edit incorrect data
• Identify Overdue Payments
• Track recurring bills

Project Structure

AI-Invoice-Processing/
│── app.py                   # Main script to process invoices
│── config.py                # Configuration settings
│── database.py              # MongoDB integration
│── email_client.py          # Fetches emails via IMAP
│── invoice_processor.py     # Extracts invoice data from PDFs and texts
│── dashboard.py             # Flask web dashboard
│── static/
│   └── styles.css           # Styles for UI
│── templates/
│   ├── dashboard.html       # Invoice list page
│   ├── edit.html            # Edit invoice details
│── .env                     # Environment variables
│── requirements.txt         # Python dependencies
│── README.md                # Documentation

Example Usage

This section demonstrates how the AI Supported Invoice Processing Tool works with real examples.

---

Sample Invoices for Testing

To test the tool, attach one of the sample invoices to an email and process it.

• Standard Invoice (USD): static/invoices/Invoice_Details_1.pdf
• UK Invoice with VAT (GBP): static/invoices/Invoice_Details_1.pdf

How to use these sample invoices:

1. Send an email to yourself with the subject "Invoice INV-2025-00123".
2. Attach one of the invoice PDFs from the paths mentioned above.
3. Run the invoice processor script:

   python app.py

4. Start the web dashboard and view the extracted invoice:

   python dashboard.py

Screenshots: How it looks in the Dashboard

Invoice Management Dashboard

Once an invoice is processed, it will be displayed in the web dashboard.

• Columns:
  • Sender, Invoice Number, Amount, Due Date, Processed At, Status
• Invoice Status:
  • Recurring: Automatically detected repeating invoices.
  • Overdue: Past-due invoices.
  • No Due Date: Invoices without a due date.
  • No Status: Default state for invoices without special condition
• Actions:
  • Edit: Modify invoice details.

Dashboard Screenshot:

Editing an Invoice

This page allows users to edit an existing invoice.

• Editable Fields:
  • Invoice Number
  • Amount
  • Due Date
• Action Buttons:
  • "Save Changes" – Updates the invoice.

Edit Invoice Screenshot: