Inspiration
We are a group of college students faced with limited budgets. We have found that we tend to lose a lot of our money on very small expenses that add up, and sometimes have trouble logging our expenditures because it takes a lot of time and we find it rather tedious.
What it does
We developed an app where users can establish their budget goals spread out along several categories: food, drinks, subscriptions, small payments, transport, and others. Then, users can upload a picture of their receipts from their photos and let AI classify the items into the categories, and keep track of how much they are spending per category. In their dashboard, the users can see what they established in their budget plan, and how much they have currently spent in that category to have an immediate visual aid to guide their spending. To make the app more engaging, we made it so that users start out with an ant that drops two leaves per day. For every week that they meet their budgeting goals, they receive another ant to earn more money. If the user opens the app on a new day, they can collect the leaves that their ants drop and receive a bonus for every day that they log in.
How we built it
Smart Receipt Scanner
• Snap a photo of any receipt. • AI-powered OCR extracts store, amount, date, and category. • Automatic expense logging in seconds.
Intelligent Expense Tracking
• 6 predefined categories: Food, Drinks, Subscriptions, Small Payments, Transport, and Others. • Weekly expense summaries with visual charts. • Real-time budget monitoring per category.
Gamification System
• Collect "ants" (workers) as saving milestones. • Earn "leaves" (virtual currency) by staying under budget. • Use these leaves to gain prizes and more ants. • Daily login bonuses that grow with consistency.
Personalized Goals
• Set spending limits for each category. • Track progress with visual indicators. • Get notified when approaching limits.
Technologies Used
Backend (REST API)
• Python 3.11 + Flask • MySQL database hosted on Aiven Cloud • Google Gemini 2.5 Flash for AI-powered receipt analysis • Swagger/Flasgger for API documentation • python-dotenv for secure credential management
Frontend (iOS App)
• Swift + SwiftUI • URLSession for API communication • Vision framework for camera integration • Charts framework for expense visualizations
Infrastructure
• Aiven Cloud Database (MySQL) • Git + GitHub for version control • Flasgger Client for API testing
AI/ML
• Google Generative AI (Gemini 2.5 Flash) • Custom prompt engineering for receipt OCR • Intelligent category classification
Challenges we ran into
The main challenge we encountered is that most of us had not previously participated in a hackathon, so we did not really know how to organize ourselves properly and wanted to build something that exceeded our capabilities given the time constraint. As such, we could not implement some of our functionalities.
Accomplishments that we're proud of
We are proud that we managed to build a functional app in such a short period of time with a database, API, and integrated API calls into our Swift application, something that we had not done previously. We really like our idea, and would love to take it further given how useful it could prove to be for many college students and even a broader audience who is looking to improve their spending habits.
What we learned
We learned that we must be better at establishing our key tasks and formulating a proposal that we could actually build given the limited work time. We also learned how to use many technologies, such as Python for APIs, connecting it to Swift, and how to engineer a prompt such that GeminiAI can analyze the receipts to facilitate the user's tracking experience.
What's next for Ant Farm
Future Roadmap
• Implement security • Improve existing functions • Allow users to win customized ants based on the categories that they meet their goals • AI-powered spending predictions. • Integration with bank APIs for automatic expense import.
Log in or sign up for Devpost to join the conversation.