InvoSync

InvoSync is an AI-powered invoice and purchase order reconciliation system that uses Tesseract OCR and RapidFuzz to detect, extract, and compare data from scanned invoices and purchase orders. It automatically highlights mismatches, corrects inconsistencies, and exports clean, standardized results — saving hours of manual verification work.

---

Features

• OCR Extraction: Reads invoice and purchase order data using Tesseract OCR.
• Discrepancy Detection: Automatically compares and corrects vendor, quantity, and pricing mismatches.
• Data Export: Generates corrected CSV reports and discrepancy summaries.
• Modern Frontend: Built using React.js + Vite for a fast and smooth UI.
• Seamless Backend Integration: Flask backend for OCR, comparison, and API handling.
• Modular Architecture: Clean separation of OCR, parsing, comparison, and export logic.

---

Project Demo

---

Tech Stack

Layer	Technologies
Frontend	React.js, Vite, TailwindCSS
Backend	Python, Flask
OCR Engine	Tesseract OCR
Text Matching	RapidFuzz
Data Processing	Pandas, Regex
File Handling	Pillow, pdf2image

---

Project Structure

InvoSync/
│
├── backend/
│   ├── app.py               # Flask app entry point
│   ├── compare.py           # Handles comparison between PO and invoice
│   ├── export.py            # Exports corrected data to CSV
│   ├── ocr.py               # OCR logic using Tesseract
│   ├── parse.py             # Parses extracted text into structured format
│   └── requirements.txt     # Backend dependencies
│
├── frontend/
│   ├── index.html
│   ├── package.json
│   ├── package-lock.json
│   └── src/
│       ├── App.jsx
│       ├── main.jsx
│       ├── components/
│       │   ├── HowItWorks.jsx
│       │   ├── Page.jsx
│       │   └── Shell.jsx
│       ├── lib/
│       │   └── api.js
│       └── pages/
│           ├── Dashboard.jsx
│           ├── Exports.jsx
│           ├── Login.jsx
│           ├── RecordDetail.jsx
│           ├── Records.jsx
│           ├── Review.jsx
│           ├── Settings.jsx
│           ├── Signup.jsx
│           └── Upload.jsx
│
└── README.md

---

Architecture Diagram

     +---------------------+
     |  Frontend (React)   |
     |  ────────────────   |
     |  Upload Invoices     |
     |  View Discrepancies  |
     +----------+----------+
                |
                v
     +---------------------+
     |   Backend (Flask)   |
     |   ───────────────   |
     | OCR Extraction       |
     | Parsing & Comparison |
     | Export to CSV        |
     +----------+----------+
                |
                v
     +---------------------+
     |   Tesseract OCR     |
     |  Image → Text       |
     +---------------------+

---

Setup & Deployment

Backend Setup

cd backend
python -m venv venv
source venv/bin/activate      # (Windows: venv\Scripts�ctivate)
pip install -r requirements.txt
python app.py

Backend runs on http://localhost:5000

---

Frontend Setup

cd frontend
npm install
npm run dev

Frontend runs on http://localhost:5173

---

How Tesseract OCR is Used

Tesseract OCR is an open-source engine developed by Google that extracts text from images and PDFs.
In InvoSync, it’s used inside backend/ocr.py to process invoice and purchase order files.

Pipeline:

1. Preprocess image using Pillow and ImageEnhance for better clarity.
2. Extract text via:

   import pytesseract
   text = pytesseract.image_to_string(image)

3. Parse the extracted text into structured fields (item name, quantity, price, total).
4. Compare extracted invoice and purchase order details using RapidFuzz for fuzzy string matching.

---

Workflow

1. Upload invoice.jpg and purchase_order.jpg through the frontend.
2. Backend extracts and parses data using Tesseract OCR.
3. Mismatches (e.g., price/quantity/vendor) are identified.
4. Corrected invoice and discrepancy report are exported as:
   • corrected_invoice.csv
   • discrepancy_report.csv

---

Output Files

File	Description
corrected_invoice.csv	Final corrected invoice after comparison
discrepancy_report.csv	Log of mismatches and corrections made

---

Contributors

Developed by:
Ashish Kumar Nanda
Kushagra Gupta
Rashi Gupta
Radhi Pahuja

---

If you find this project useful, please consider giving it a ⭐ on GitHub!