DEV Community

softGpt 9299
softGpt 9299

Posted on

SplitEase - I Built an AI-Powered Bill Splitter So My Roommates Would Stop Owing Me Money

This is a submission for the MLH x DEV Writing Challenge

What I Built

SplitEase (live at split.elitecrew.online) is a zero-paywall expense-splitting platform for roommates, friend groups, and trips — built to kill the awkward "I'll pay you back later" conversation for good.

The idea came from a very real, very annoying problem: tracking hostel bills, flat rent, and trip expenses across a WhatsApp group and a mental spreadsheet that nobody actually updates. SplitEase replaces that chaos with one place where every group tracks who paid, who owes, and how to settle up with the fewest possible transfers.

Core features:

  • Groups for everything — unlimited groups and members, with roommate, trip, and one-off "Quick Split" modes
  • Flexible splitting — equal, percentage, shares, or exact-amount splits
  • QR, link & email invites — get a group started in seconds, no signup friction
  • Real-time group chat — payments and settlements are recorded directly inside the conversation, so there's no "wait, did you actually pay?"
  • Minimum-transfer settlement suggestions — the app does the math to simplify 18 messy bills into just 2 transfers
  • Receipt photos + a built-in AI assistant — attach a receipt and let the assistant help itemize it
  • Live balance dashboard & spending charts — see exactly where the group's money went, by category, by month
  • Secure OTP login, with TLS in transit and encryption at rest (security practices inspired by OWASP)

It's completely free — no per-member fees, no trial period, no feature gates — and it runs both as a web app and as a mobile app on Google Play.

Demo

🔗 Live app: https://split.elitecrew.online

📱 Also on Google Play (React Native app for Android & iOS)

Partner Technologies

SplitEase is a full-stack build end to end:

  • Frontend: Next.js, React, TypeScript, Tailwind CSS, Framer Motion, Recharts, Axios
  • Backend: Node.js + Express, MongoDB (via Mongoose), JWT authentication, Socket.io for real-time chat and live balance updates
  • Infrastructure: Firebase, Cloudinary for receipt/image storage, Nodemailer for email delivery
  • AI & OCR: OpenAI and Google Gemini power the in-app assistant, with Tesseract.js handling optical character recognition on receipt photos — together they turn a photographed receipt into itemized, split-ready expense data
  • Mobile: React Native (Android & iOS), Firebase
  • Quality: Jest test suites, CodeQL static analysis, and CI running fast-check property-based security tests on every push

Hackathon Experience

Building SplitEase at an MLH hackathon was less about chasing a trendy idea and more about finally fixing a problem I'd been living with for years — the awkward "who owes who" math that follows every trip, every shared flat, every group dinner. Hackathon weekends have a very specific energy: a hard deadline, a working demo as the only acceptable outcome, and no time to overthink the stack. That pressure is exactly what turned "I should build an expense splitter someday" into a real, deployed product with real-time chat, AI-assisted receipts, and a live settlement engine in one weekend.

The biggest takeaway was how much a tight hackathon timebox forces good scoping — deciding what's a "coming soon" feature (multi-currency, CSV export, UPI payouts) versus what has to work flawlessly on demo day (splitting, settling, chatting). That discipline is still shaping how I'm building out SplitEase today.

Top comments (0)