Invoice Intelligence Extractor

AI-powered invoice data extraction. Upload any PDF invoice and get structured, validated data in seconds — no manual entry required.

Built with Google Gemini and Python. Designed for operations teams, accounting firms, and logistics companies that process high volumes of invoices.

---

What it does

• Extracts vendor name, invoice number, dates, line items, subtotal, tax, and total
• Assigns confidence scores to every extracted field
• Exports results to CSV or JSON with one click
• Handles digital PDFs in any language
• Processes an invoice in under 3 seconds

Before and after

Before: An accounts team manually types 200 invoices into a spreadsheet — 3 hours of work per day.

After: Upload the PDFs, download the CSV — 4 minutes.

---

Setup

git clone https://github.com/YOUR_USERNAME/invoice-extractor
cd invoice-extractor
python3 -m venv venv && source venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
# Add your GEMINI_API_KEY to .env

Get a free Gemini API key at aistudio.google.com

---

Run

# Smoke test (confirms API connection works)
python tests/test_extractor.py

# Launch the web app
streamlit run app.py

---

Project structure

invoice-extractor/
├── app.py                  # Streamlit web UI
├── config.py               # Centralised configuration
├── core/
│   ├── gemini_client.py    # Gemini API wrapper with retry logic
│   ├── extractor.py        # Extraction orchestration engine
│   └── models.py           # Typed data models
├── utils/
│   ├── pdf_parser.py       # PDF → text conversion
│   └── export.py           # CSV / JSON export
└── tests/
    └── test_extractor.py   # Smoke test

---

Tech stack

• AI: Google Gemini 1.5 Flash
• UI: Streamlit
• PDF parsing: PyMuPDF
• Language: Python 3.11+

---

Built by

Robotics & AI Engineering student specialising in Python automation and AI integration.

Available for custom AI automation projects — [tiwarianshika09.05.2005@gmail.com]