BorrowEasy

GitHub Repository

Inspiration

As college students, we’ve experienced firsthand how confusing and manipulative loan offers can be especially when applying for education loans. From hidden fees to emotionally charged offers that pressure you into quick decisions, we’ve felt the trap close in real time. That personal frustration, combined with India’s growing financial fraud crisis, is what led us to create BorrowEasy.

In 2023, India saw over 1.13 million cyber financial fraud cases, resulting in losses of more than ₹7,488.6 crore. Loan scams, fake apps, phishing links, hidden clauses, bait-and-switch tactics are a major contributor to this crisis. Traditional financial tools only assess numeric terms, but none of them evaluate the cognitive and emotional manipulation built into the loan offer itself. BorrowEasy was built to fill this critical gap.


What it does

BorrowEasy is a cognitive loan trap detector and personal financial safety tool designed to help users make informed, emotionally neutral, and financially sound borrowing decisions. It analyzes both the terms and the presentation of loan offers to flag manipulative practices and guide users toward safer alternatives.

Key Features:

  • Cognitive Loan Trap Detection
    Detects hidden fees, vague repayment terms, exorbitant interest rates, urgency-inducing language, and emotionally manipulative design patterns using NLP and behavioral economics principles.

  • Risk Flagging & Explanation
    Clearly explains red flags in simple terms and shows users why a loan might be dangerous — not just technically, but psychologically.

  • Financial Metric Calculator
    Calculates key metrics like APR, EMI, total repayment, and prepayment penalties so users know exactly what they’re getting into.

  • Loan Recommender
    Based on your input (e.g., income, tenure preference, risk appetite, and purpose), BorrowEasy recommends safer, transparent loan options from trusted providers, helping users not just avoid traps, but find better alternatives.

  • Education Hub
    A built-in resource center offering simple, well-structured modules to teach users about:

    • How to read loan documents
    • Red flags and scam patterns
    • Psychological manipulation in fintech
    • Real-world scam case studies and how they could have been avoided
  • Future Finance Predictor
    A forecasting tool that analyzes your current loans, spending behavior, and projected needs to predict your financial health. It warns you about:

    • Potential over-indebtedness
    • Upcoming repayment stress
    • Whether you can safely take another loan

How we built it

Due to time constraints, we used Gemini models to handle natural language processing and emotional analysis of loan texts.

System Architecture:

  • Frontend:
    A web interface built using React and HTML/CSS where users can input loan terms or upload documents.

  • Backend:

    • Gemini Models: Used for fast and powerful NLP parsing of loan clauses, detection of manipulative language, and content classification.
    • Risk Assessment Engine: Applied financial logic (APR, EMI, penalties) to identify red flags.
    • Knowledge Base: A repository of predatory loan patterns and scam examples.
    • Recommendation Engine: Suggests safer, well-structured loan options based on user profile.
    • Prediction Module: Forecasts future financial risk using current and projected borrowing data.
  • Tech Stack:

    • Gemini for NLP and AI
    • Python (Flask/FastAPI)
    • React, HTML, CSS for the UI

Challenges we ran into

  • Defining “manipulation” in a way that can be detected programmatically
  • Parsing unstructured, diverse loan documents using NLP
  • Balancing technical accuracy with simplicity for users unfamiliar with finance
  • Working with limited time, which led us to integrate Gemini instead of training custom models

Accomplishments that we're proud of

  • Built a fully functional MVP that integrates financial, behavioral, and predictive tools
  • Successfully flagged predatory elements in real-world loan samples during testing
  • Turned our own negative experiences into a product that can protect thousands of other borrowers
  • Bridged the gap between tech, finance, and psychology in a real-world application

What we learned

  • Scams today don’t always look like scams, they’re subtle, persuasive, and often legal-looking
  • Emotional manipulation is just as harmful as financial exploitation in digital lending
  • Gemini’s NLP capabilities allowed rapid prototyping and helped us focus on logic over infrastructure
  • Borrowers need more education and transparency, especially in fast-growing markets like India

What's next for BorrowEasy

  • Train our own transformer models for deeper manipulation detection
  • Partner with banks, credit platforms, and fintech regulators
  • Add regional language support and voice input for wider accessibility
  • Integrate browser/mobile plugins to analyze loan terms in real-time
  • Launch an awareness campaign #ThinkBeforeYouSign to educate users about cognitive loan traps

Built with ❤️ by students who were once trapped, and now want to make sure others never are.

Built With

Share this project:

Updates