Nexpay

Intelligent Commerce. Profitable Decisions. Seamless Payments.

Comment

PAYPAL NEXUS

Think Smarter. Pay Smarter. Save Bigger.

Inspiration

Online shoppers often spend considerable time searching across multiple platforms to find the best price for a product. The same product can have different prices across merchants, and comparing listings manually can be confusing and time-consuming.

We wanted to make smarter shopping decisions easier.

That's why we built PAYPAL NEXUS — an AI-powered shopping intelligence platform designed to help users discover product offers, compare prices, identify potential savings, and make more informed purchasing decisions.

Our vision is simple: shopping should not just be about finding a product; it should be about finding its value.

What it does

PAYPAL NEXUS combines AI assistance, external shopping search, and deterministic product-matching logic to create a smarter price-comparison experience.

Key features include:

  • Product Discovery: Browse a curated product catalogue with product details and reference prices.
  • AI-Powered Product Understanding: Use Groq-powered AI to assist with product interpretation and comparison explanations.
  • Shopping Search: Retrieve product listings from Google Shopping through the SerpApi integration.
  • Intelligent Product Matching: Validate external listings against product brand and model information to reduce incorrect comparisons.
  • Price Comparison: Compare verified matching offers against the NEXUS reference price.
  • Savings Calculation: Calculate potential savings and savings percentages using Python-based logic.
  • Transparent Results: Display merchants, product links, search information, and available listing details.
  • Demo Checkout: Demonstrate a simulated shopping journey without processing real payments.

How we built it

We built PAYPAL NEXUS using a full-stack architecture with a React frontend and Python backend.

Frontend

  • React.js
  • Vite
  • JavaScript
  • CSS

Backend

  • Python
  • FastAPI
  • SQLAlchemy
  • REST APIs

Database

  • PostgreSQL
  • JSONB for flexible product specifications

AI and External Search

  • Groq API for AI-powered product interpretation and explanations
  • SerpApi Google Shopping API for external shopping listings

Core Intelligence

  • Python-based product matching and validation
  • Product identity and model checks
  • Price comparison and savings calculations
  • Verified-match filtering before calculating savings

The workflow begins when a user selects a product. The backend retrieves its reference information, performs an external shopping search, evaluates the returned listings, and identifies products that meet the matching criteria. Only verified matches are considered by the price-comparison engine.

Groq helps explain the comparison results, while Python remains responsible for validating product matches and calculating prices and savings.

Challenges we faced

One of our biggest challenges was accurate product matching.

A shopping search for a specific product can return similar-looking products, different models, or variations from the same brand. For example, searching for boAt Nirvana Air Buds may return Nirvana Ion, Nirvana Crown, or other products.

Comparing these products directly could produce misleading savings.

To address this, we developed a stricter product-matching approach that checks brand and distinctive model information before accepting a listing. Unmatched listings are rejected rather than being used to calculate savings.

We also worked on separating AI-generated explanations from authoritative calculations. An AI model should not invent product prices or claim that an unverified listing is a valid match. Therefore, our Python comparison engine is the source of truth for matching status and savings calculations.

Integrating the frontend, backend, PostgreSQL database, AI service, and external shopping search within a limited hackathon timeline was another major challenge.

What we learned

Building PAYPAL NEXUS helped us understand how to combine AI models with real-world APIs and deterministic backend logic.

We learned:

  • How to integrate LLMs into practical applications.
  • How to connect external search APIs with FastAPI.
  • Why product identity validation matters in shopping applications.
  • How to design reliable price-comparison logic.
  • How to work with PostgreSQL and SQLAlchemy.
  • How to build and integrate a React frontend with REST APIs.
  • Why transparency and data validation are essential when building AI-powered financial and shopping experiences.

What's next

We envision expanding PAYPAL NEXUS with:

  • More shopping sources and merchant integrations.
  • Improved product matching with verified model aliases.
  • Historical price tracking.
  • Price-drop alerts and watchlists.
  • Personalized shopping recommendations.
  • More comprehensive merchant, delivery, and availability information.
  • A more advanced product value-scoring system.

Our vision

PAYPAL NEXUS is more than a price-comparison interface. It is our step toward making online shopping more intelligent, transparent, and value-driven.

Think Smarter. Pay Smarter. Save Bigger.

Note: This hackathon prototype uses external shopping search results and a simulated checkout. It does not process real payments or guarantee live merchant inventory or prices.

Built With

Share this project:

Updates

Submission history