The problem Wallet Wizard solves
A native mobile application that helps users manage their finances through automated expense tracking, smart budgeting, and AI-powered financial insights. The app integrates LLAMA, a large language model, for conversational queries and personalized financial advice.
Key Features
- Automated Expense Tracking
Users can connect their bank accounts (via Plaid or mock data) to automatically track their income and expenses.
The backend categorizes expenses using predefined categories (e.g., groceries, bills, entertainment) based on transaction data.
The AI learns user spending patterns and improves the accuracy of categorization over time.
- Smart Budget Creation
The app generates personalized budgets for users based on historical spending patterns and their income.
Users can set adjustable budgets for each category (e.g., groceries, bills, dining).
Dynamic budget suggestions adjust to changes in financial situations (e.g., salary changes).
- Spending Insights & Alerts
Detailed insights into spending trends: top spending categories, average weekly spending, and irregular spending spikes.
The app sends real-time alerts when users approach or exceed their budget in specific categories.
- Saving Goals
Users can set financial goals (e.g., saving for a vacation, building an emergency fund).
The AI provides suggestions on how much to save each month based on income, expenses, and goals.
Progress toward savings goals is visually displayed with charts.
- Expense Forecasting
The app uses historical data to predict future expenses and cash flow.
Notifications remind users of upcoming payments and suggest potential savings.
- LLAMA-Powered Conversational Assistance
Users can ask natural language questions like:
“How much did I spend on dining last week?”
“What’s my top spending category this month?”
“Can I afford to buy a $300 item?”
LLAMA interprets these queries and provides real-time answers based on the user's financial data.
Challenges we ran into
Eas for production level check was a new thing for us. Learnt and implemented it within some hours.
One problem which still remains is making the apk using Eas since project is large due to the use of LLAMA and cannot be solved without buying EAS premium cloud