Skip to content

Repository files navigation

LoanTruth Banner

LoanTruth · See Through the Interest-Rate Illusion

Upload a loan screenshot. AI rips off the financial product's mask and shows you the real cost.

License: MIT React TypeScript Vite Zhipu GLM-4V-Flash PRs Welcome

🌍 Helping young people worldwide see through predatory lending — one person at a time.


Why this exists

Billions of young people are having their lives destroyed by high-interest loans.

They're not stupid — the interest rates are deliberately obfuscated.

"Only 0.05% daily" sounds cheap. "0.6% monthly fee" looks low. But math doesn't lie.

LoanTruth's mission, in one line:

Use math and AI to expose predatory financial products. Help whoever we can.


🚀 Live Demo

👉 Try it now — fast, global CDN

(Mirror: GitHub Pages)

LoanTruth demo

Upload a screenshot → AI extracts the terms → Halley IRR verifies the real APR. "0.05% daily" ≈ 18.25% nominal, but 30.92% real after upfront fees.


What it does

Feature Description
📸 Screenshot upload Multi-image stitching (homepage ad + repayment schedule)
🤖 AI recognition Zhipu GLM-4V-Flash extracts principal, term, hidden fees
🧮 Algorithmic verification Halley-method IRR computed locally (3rd-order convergence)
🚨 Trap detection Upfront interest, forced insurance, early-repayment penalties
💰 Debt-to-income ratio Enter monthly income, see how close to collapse you are
Debt-cycle warning Borrowing to repay debt = slow suicide
🎲 Investment warning Don't gamble your way out of debt — 90% lose more
🛟 Survival roadmap Concrete, actionable steps out of debt

Quick Start

# 1. Clone
git clone https://github.com/Focus688/LoanTruth.git
cd LoanTruth

# 2. Install dependencies
npm install

# 3. Configure the Zhipu GLM-4V-Flash API key
#    Get a free key at https://bigmodel.cn
echo "VITE_ZHIPU_API_KEY=your_key" > .env.local

# 4. Run
npm run dev

🆓 GLM-4V-Flash has a free tier — enough for personal use. The core IRR math works without any API key.


Tech Stack

Layer Choice
Framework React 19 + TypeScript
Build Vite 6
Styling Tailwind CSS
AI Zhipu GLM-4V-Flash (free tier)
Algorithm Halley IRR (3rd-order convergence, client-side)
Charts Recharts
Icons Lucide React

Architecture

Upload screenshot
        ↓
Zhipu GLM-4V-Flash extracts (principal / term / payment / fees)
        ↓
Local Halley IRR → precise real APR
        ↓
Double verification → AI estimate vs. algorithm
        ↓
Report → risk level + traps + burden analysis + survival roadmap

Why client-side? User data never hits our servers. Images go directly to the AI API; computation happens in the browser. Income data lives only in memory and disappears on refresh.


Risk Tiers

Tier Real APR Meaning
🟢 Low < 10% Normal bank loans, mortgages
🟡 Medium 10–24% Credit cards, consumer loans
🟠 High 24–36% High-interest cash loans — pay off ASAP
🔴 Predatory > 36% Usury — legally unprotected in most jurisdictions

Why contribute

This isn't a "fun project."

  • You may know someone trapped in a debt cycle right now
  • You may have seen young people ruined by predatory loans
  • You may have been burned by "0.05% daily" yourself

Every person this project helps is one fewer person harvested by the financial machine.

How you can help:

  • 🎨 Design — make the UI more impactful and approachable
  • 🧮 Algorithm — better IRR, more financial products
  • 🌍 Translation — reach young people in more countries
  • 📢 Spread the word — help whoever we can

Contribution guide →


Roadmap

  • MVP — single upload + AI analysis + IRR verification
  • Debt-to-income ratio
  • Debt-cycle warning
  • Investment/gambling warning
  • Survival roadmap
  • Multi-language (EN/ES/JP/ID)
  • More product templates (auto / mortgage / credit card)
  • Local OCR fallback (works without any API)
  • PWA / Telegram Bot

Math doesn't lie. See it, and you win.

Try it · Report issue · Contribute

Made with ❤️ for every young person trapped in debt.

About

Help global "retail investors" break free from the control of financial loans

Topics

Resources

Contributing

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages