CardWise

CardWise - More Rewards, Less Effort

Comment 5

Inspiration

We've all experienced that moment at checkout: staring at multiple credit cards in our wallet, wondering which one to use. Most people just grab their default card, unknowingly leaving hundreds or even thousands of dollars in rewards on the table every year.

The problem isn't lack of awareness, it's lack of real-time help. Existing credit card comparison tools require you to manually research cards before shopping, and they don't help at the actual moment of purchase when you need it most.

We built CardWise to be your financial co-pilot: an ecosystem that detects when you're checking out and instantly tells you which card maximizes your rewards for that specific purchase.

What it does

CardWise is a real-time credit card rewards optimizer that combines a Chrome extension with a web dashboard to help users maximize their credit card rewards effortlessly.

Core Features

  1. Intelligent Checkout Detection

    • Automatically detects when you're about to make a purchase on any e-commerce site
    • Analyzes URLs, form fields, and page content to trigger at the right moment
    • Works across thousands of online merchants

  2. Instant Card Recommendations

    • Displays a side panel showing your optimal card for the current purchase
    • Analyzes 200+ credit cards in our database to find the best match
    • Considers multiple factors:
      1. Category-specific rewards (5x on travel, 3x on dining, etc.)
      2. Merchant-specific bonuses (extra points at airlines, hotels, gas stations)
      3. Point value multipliers
      4. Annual fees and APR
      5. Your income level and risk tolerance
      6. Your financial goals (travel rewards, cashback, credit building)

  3. Risk-Aware Recommendations

    • Never suggests cards with fees you can't afford (income-based filtering)
    • Avoids high-APR cards for users who carry balances
    • Tailored suggestions based on your credit profile

  4. Goal-Driven Optimization

    • Optimize for travel points, cashback, or credit building
    • Track progress toward rewards milestones
    • Get recommendations for new cards that fit your spending patterns

  5. Web Dashboard

    • Manage your credit card wallet
    • Track "money left on the table" analytics
    • See personalized card recommendations based on your spending history

How we built it

Data Collection and Processing We automated web scraping of 10 major bank websites (Capital One, Chase, Amex, Citi, Bank of America, Wells Fargo, US Bank, Discover, Synchrony, Barclays) and extracted 200+ credit cards with structured data including annual fees, APR, rewards categories, bonus multipliers, and merchant-specific bonuses.

Merchant Intelligence System We integrated a Merchant Category Code (MCC) database based on ISO 18245 standards and created a hybrid merchant-to-category mapping system using:

  • Static MCC code library (982 codes)
  • AwardWallet merchant lookup data
  • Keyword-based detection (URLs, merchant names)
  • AI fallback for unknown merchants

Chrome Extension Architecture

  • Content scripts for checkout detection and page analysis
  • Background service worker for API communication
  • Side panel UI for seamless user experience
  • Cross-origin request handling with proper security

Recommendation Engine Our scoring algorithm combines multiple factors: Card Score = (Category Rewards * Point Value) - Annual Fee + Merchant Bonus + Risk Adjustment

The engine works in layers:

  1. Merchant Detection: Maps URLs to categories (Delta.com → airlines → travel)
  2. Category Matching: Finds cards with best rewards for that category
  3. Risk Filtering: Removes cards outside user's financial comfort zone
  4. Goal Optimization: Prioritizes cards aligned with user objectives
  5. AI Enhancement: Google Gemini API provides intelligent ranking
  6. Fallback System: Rule-based scoring ensures 100% uptime when AI is unavailable

Key Technical Implementation Checkout Detection Algorithm:

//Weighted scoring across multiple signals
const checkoutIndicators = {
  url: /checkout|cart|payment|billing/i,
  forms: hasPaymentFields(),
  buttons: /place order|pay now|complete purchase/i,
  ssl: location.protocol === 'https:'
};

Merchant Categorization

//Intelligent keyword mapping
const merchantCategories = {
  'delta|united|airline': 'travel',
  'hilton|marriott|hotel': 'travel',
  'whole foods|kroger|grocery': 'groceries',
  'shell|chevron|exxon': 'gas'
};

Challenges we ran into

Merchant Categorization Complexity The Problem: How do you map "Delta.com" to "this is an airline purchase, so use your Chase Sapphire Reserve with 5x travel rewards"? The Solution: We built a comprehensive keyword mapping system with fallback logic:

  • Primary: Check URL for airline/hotel/merchant keywords
  • Secondary: Analyze merchant name from payment forms
  • Tertiary: Look at page content and meta tags
  • Fallback: Use generic category-based recommendations

Checkout Detection Accuracy The Problem: Not all e-commerce sites structure their checkout the same way. Some use "/cart", others "/checkout", some have multi-step processes. The Solution: We developed a weighted scoring system that combines multiple signals:

  • URL patterns (30% weight)
  • Form field analysis (25% weight)
  • Button text detection (20% weight)
  • SSL verification (15% weight)
  • Page title keywords (10% weight) Only trigger when confidence score > 70%.

What's next for CardWise

Mobile Application Build iOS and Android apps with in-store recommendations via NFC and GPS location detection. For example, when you're at Target, the app notifies you to use Chase Freedom for 5% back this quarter.

Spending Analytics Dashboard Track what you're earning versus what you're leaving on the table. Show insights like "You could have earned $243 more last month by optimizing your card usage."

Real Bank Integrations Connect via Plaid or Yodlee for automatic transaction import, eliminating manual wallet management by automatically detecting your cards.

Goal Tracking System Provide visual progress toward rewards milestones, such as "You're 8,000 points away from a free flight to Europe."

Smart Alert System Send proactive notifications about rotating bonus categories, like "Q2 bonus categories are live: 5% on gas with Discover It."

Machine Learning Card Recommendations Analyze spending patterns to recommend new card acquisitions. For example, "Based on your $600/month grocery spending, you should get the Amex Gold."

International Expansion Support credit cards from Canada, UK, and EU markets.

Mission Statement Our mission is to democratize financial optimization. Everyone should get the rewards they deserve, regardless of financial expertise.

Built With

+ 5 more
Share this project:

Updates