AI Expense Tracker

A full-stack expense tracking app that uses AI to parse natural language input.

Built by: Anglesvar Appiya Chandrasekar GitHub: https://github.com/Anglesvar Time to build: [3 hours] (with AI assistance)

🎥 Demo

React-Native-Android NodeJS

🛠️ Tech Stack

• Mobile: React Native, Expo, TypeScript
• Backend: Node.js, Express, TypeScript
• Database: SQLite
• AI: [Groq/OpenAI/Gemini] API

🚀 Setup Instructions

Prerequisites

• Node.js 18+
• npm or yarn
• Expo CLI
• [AI Service] API key

Backend

cd backend
npm install
cp .env.example .env
# Add your AI API key to .env
npm run dev

Mobile

cd mobile
npm install
npm start
# Scan QR code with Expo Go app

📁 Project Structure

ai-expense-tracker/
    backend/
    ├── src/
    │   ├── index.ts                        # Application entry point
    │   ├── config/                         # Configuration files
    │   │   └── config.ts                   # Environment config
    │   ├── database/                       # Database layer
    │   │   ├── connection.ts               # DB connection & initialization
    │   |   ├── utils.ts                    # DB utility functions
    │   │   └── dto/                        # Data Transfer Objects
    │   │       ├── create-expense.dto.ts
    │   │       ├── update-expense.dto.ts
    │   │       └── index.ts
    |   | 
    │   ├── services/                       # Business logic layer
    │   │   └── expense.service.ts
    │   │   └── expense-parser.service.ts
    │   ├── types/                          # Business logic layer
    │   │   ├── expense.ts
    │   │   ├── expense-filters.ts
    │   │   ├── parsed-expense.ts
    │   │   └── index.ts
    │   ├── routes/                         # API routes layer
    │   │   ├── index.ts                    # Routes aggregator
    │   │   └── expenses.routes.ts          # Expense endpoints
    │   ├── middleware/                     # Express middleware
    │   │   ├── error.middleware.ts
    │   │   └── validation.middleware.ts
    │   └── utils/                          # Helper utilities
    │       └── async-handler.ts
    │       └── validate-and-normalize.ts
    ├── .env                                # Environment variables
    ├── .env.example                        # Example env file
    ├── package.json
    └── tsconfig.json

    mobile/
    ├── App.tsx                             # Main app entry point
    ├── screens/
    │   └── ExpenseTrackerScreen.tsx        # Main screen with business logic
    ├── components/
    │   ├── index.ts                        # Component exports
    │   ├── Header.tsx                      # App header with title
    │   ├── ExpenseInput.tsx                # Input field with Add button
    │   │   ├── TextInput
    │   │   ├── AddButton
    │   │   └── LoadingIndicator
    │   ├── SuccessCard.tsx                 # Success feedback card
    │   ├── ExpenseList.tsx                 # FlatList with pull-to-refresh
    │   └── ExpenseItem.tsx                 # Individual expense item
    │       ├── CategoryIcon (emoji + name)
    │       ├── Amount
    │       ├── Description
    │       ├── TimeAgo
    │       └── DeleteButton
    ├── api/                                 # API service layer
    |       helpers/
    │           ├── expense-api-error.ts
    │           ├── fetch-with-timeout.ts
    │           └── handle-response.ts
    │       ├── add-expense.api.ts
    │       ├── delete-expense.api.ts
    │       ├── get-expense.api.ts
    │       └── index.ts
    ├── types.ts                            # TypeScript interfaces
    ├── utils.ts                            # Utility functions
    ├── config.ts                           # Configuration
    ├── package.json                        # Dependencies
    └── tsconfig.json                       # TypeScript config

    |-- README.md

🤖 AI Prompt Design

I used this system prompt for expense parsing:

    You are an expense parser. Extract expense information from natural language input.

    RULES:
    1. Extract the amount as a number (no currency symbols)
    2. Default currency is INR unless explicitly mentioned (USD, EUR, etc.)
    3. Categorize into EXACTLY one of these categories:
    - Food & Dining (restaurants, cafes, food delivery, groceries)
    - Transport (uber, ola, taxi, fuel, parking, metro)
    - Shopping (clothes, electronics, amazon, flipkart)
    - Entertainment (movies, netflix, spotify, games)
    - Bills & Utilities (electricity, water, internet, phone)
    - Health (medicine, doctor, gym, pharmacy)
    - Travel (flights, hotels, trips)
    - Other (anything that doesn't fit above)
    4. Description should be a clean summary (not the raw input)
    5. Merchant is the company/store name if mentioned, null otherwise

    RESPOND ONLY WITH VALID JSON, no other text:
    {
        "amount": <number>,
        "currency": "<string>",
        "category": "<string>",
        "description": "<string>",
        "merchant": "<string or null>"
    }

    If the input is invalid or you cannot extract an amount, respond:
    {
        "error": "Could not parse expense. Please include an amount.",
        "amount": null
    }