A Clean-MVVM Android application for splitting bills using AI-powered receipt parsing.

This app was created as a demo for the talk:

AI-Accelerated Mobile Development - From idea to build, faster than ever. Even when you are the whole team.
By Gio Scalzo at the Londroid event, NewDay, 29-5-2025.

Every git commit message in this repo is the prompt used for that commit.

The detailed steps followed to build this app are documented in project_steps.md.

Split Bill is a showcase Android app that leverages AI to make splitting restaurant bills effortless. Snap a photo of a UK receipt, let AI parse the items, edit as needed, assign costs, track payments, and keep a history of events—all in a beautiful, modern UI.

• 📷 Camera and gallery integration for receipt capture
• 🤖 AI-powered receipt parsing using OpenAI gpt-4o-mini
• 🎨 Material Design 3 UI with Jetpack Compose
• 🇬🇧 UK receipt format support with quantity handling
• ✏️ Inline editing of item names, quantities, and costs
• 🛡️ Secure API key management via environment variables
• ⚡ Loading spinner during API calls
• 🧪 Mock service for testing without API calls
• 👥 Participant management and local persistence
• ➗ Flexible split modes: equal or manual assignment
• ✅ Payment tracking and event history

1. Copy the example environment file:

   cp .env.example .env

2. Edit .env and add your OpenAI API key:

   OPENAI_API_KEY=sk-your-actual-openai-api-key-here
   

3. Get your OpenAI API key from OpenAI Platform

./gradlew assembleDebug

1. Capture or select a receipt photo using the FAB.
2. AI parses the receipt (via OpenAI gpt-4o-mini) and extracts items, prices, service charge, and total.
3. Edit the parsed list if needed—change item names, costs, or quantities.
4. Add/select participants and assign items (split equally or manually).
5. Track payments—each person marks their share as paid, with optional notes.
6. View event history—see settled/unsettled bills.

• UI Layer: Jetpack Compose, Material Design 3
• Business Logic: ViewModels, Use Cases
• Data Layer: Repository pattern, mockable services
• AI Integration: OpenAI API, structured JSON output
• Network: Retrofit, OkHttp
• Persistence: Local storage for participants and event history

• API keys are stored in .env files (excluded from git)
• Images are compressed (not resized) to balance quality and API token usage
• All network calls are asynchronous with robust error handling

• Use the mock service for rapid development:

  // In MainActivity.kt, change to:
  val receiptService: ReceiptParsingService = remember { OpenAIService(useMock = true) }

• Loading spinner and error handling are built-in for all API calls.
• All features are unit tested and follow Clean-MVVM best practices.

• The full step-by-step process used to build this app is in project_steps.md.
• Every git commit message is the prompt used for that commit.

MIT License. See LICENSE for details.