SplitMind — Smart Bill Splitter
Offline-first mobile application designed to seamlessly split group bills with smart debt simplification, receipt scanning OCR, and AI-powered financial insights.
Split bills, not friendships.
SplitMind is a modern, offline-first mobile application designed to make group expenses and bill splitting effortless. Whether it’s a weekend trip, a dinner with friends, or shared apartment costs, SplitMind tracks who paid what and calculates the simplest way for everyone to settle their debts.
The Problem
Tracking shared expenses manually often leads to confusion, awkward conversations, and complex webs of “who owes whom”. Manual calculations are error-prone, and splitting receipts item-by-item is tedious and time-consuming.
The Solution: SplitMind
SplitMind removes the friction by centralizing the data, automatically parsing receipts to save manual data entry, and using a smart algorithm to drastically reduce the number of payment transfers required to settle up.
Key Technical Features
1. Smart Debt Simplification
Uses a min-cash-flow algorithm (graph theory) to minimize the number of transactions needed to settle all debts in a group. Instead of 10 complex transfers, the algorithm finds the optimal 3-4 transactions.
2. Offline-First Architecture
Fully functional without an internet connection. Write operations are saved locally first using Isar Database and synced in the background when connectivity is restored.
3. Advanced Receipt Scanning
Integrates ML Kit OCR to automatically extract items and prices from physical receipts, allowing users to easily assign items to specific members without manual data entry.
4. AI-Powered Financial Insights
Leverages the Gemini API to provide weekly natural language summaries and spending insights (e.g., “Your group spent 65% on food this week!“).
5. Flexible Split Modes
Split expenses equally, by exact item, by percentage, or input custom amounts — accommodating every scenario.
Tech Stack
- Framework: Flutter (Mobile — Android/iOS)
- State Management: Riverpod
- Local Database: Isar Database
- Routing: GoRouter
- Architecture: Feature-First Clean Architecture
- AI & ML: Google Gemini API, ML Kit Text Recognition
UI/UX Highlights
- Built with a custom Design System featuring a sleek Light Theme with Vibrant Blue (
#3B82F6) and Teal (#14B8A6) accents. - Includes micro-interactions, smooth Lottie animations for empty states, and shimmer loading patterns.
- Clear, easily readable Analytics Dashboard featuring progress bars and an interactive leaderboard (“Who Treats the Most”).