As we explored Capital One’s “Reimagine Banking” challenge, we wanted to make financial education engaging, gamified, and interactive. Our idea grew into Copper One, a platform where kids can learn financial concepts through play, track their goals, and earn rewards while parents stay involved through a secure dashboard.
Copper One is a financial education app designed for kids and families. It blends learning modules, chores, savings goals, and a fun AI companion named Penny to create a safe, educational ecosystem for money management.
What We Learned:
- Building Copper One taught us much more than just technical skills.
- Financial literacy design requires empathy — we learned how to simplify complex concepts for children.
- Team collaboration across React, Firebase, and AI pipelines improved our understanding of modular architecture.
- User testing with kids and parents helped shape our UI/UX into something truly engaging.
We also gained insight into how AI can empower education responsibly when properly sandboxed and redacted.
Financial Learning: Curated modules from Khan Academy on saving, budgeting, debt, and investments.
Goal Setting: Kids can set savings goals, track progress, and earn “Piggy Points.”
Chores System: Parents assign chores with rewards, teaching earning and responsibility.
Penny AI Assistant: Local LLM-powered tutor (Llama 3 / Phi-3) that explains financial concepts in kid-friendly terms.
Finance Quest Game: An interactive HTML5 mini-game that teaches financial decision-making.
Analytics Dashboard: Visual breakdowns of spending, savings, and progress.
Security: PIN-protected pages, data redaction, and parent-admin roles.
We built it using:
- Frontend: React 19 + TypeScript + Tailwind CSS
- Backend: Firebase (Functions + Firestore) with Node.js + Express
- AI Assistant: Local Ollama instance running Llama 3.2 3B and Phi-3 Mini
- Learning Modules: Khan Academy API integration
- Analytics: Client-side visualizations and real-time updates
- Game Engine: Embedded HTML5/Canvas-based financial learning mini-game
The architecture included: React Frontend→Firebase API Layer→Firestore (Data)→Ollama (AI Layer)
We used React Context API for global state management, Zod for schema validation, Firebase Emulators for local testing, JSON-based AI response parsing for consistent replies from Penny and our team collaborated via Git + Firebase Hosting, testing each feature in real-time.
Challenges we ran into were:
- Integrating AI locally: Running LLMs like Llama 3.2 3B locally through Ollama was challenging due to model memory constraints and JSON parsing errors.
- Security and PIN system: Balancing usability with security for kids’ accounts required careful UX and context management.
- Data Persistence: Transitioning from localStorage prototypes to Firestore syncing while keeping app state consistent.
- Game embedding: Ensuring smooth performance and cross-browser support for the Finance Quest mini-game.
- Structured AI communication: Designing system prompts and enforcing structured JSON responses for Penny’s advice.
Accomplishments That We're Proud Of:
- Completed MVP with 12 major features, including AI, analytics, and parental monitoring.
- Kid-friendly AI assistant that gives age-appropriate financial advice.
- Integrated real educational content from Khan Academy directly into the app.
- PIN-protected system that keeps sensitive actions secure.
- Gamified financial learning, making serious topics like budgeting fun.
- Fully deployable Firebase stack — ready for hosting and production use.
What's Next for Copper One:
The MVP is complete and production-ready — but we’re just getting started. Upcoming milestones:
- Firebase Authentication for cloud-based user management.
- Mobile App (React Native) for better accessibility.
- Global learning expansion with multilingual support.
- Parent insights dashboard with advanced analytics and recommendations.
- Expanded AI personalities — Penny’s friends who specialize in investing, saving, and entrepreneurship.
- Physical Piggy Bank Hardware Extension: Using RFID or sensors to detect real money deposits linked to the app.
We didn’t just build a banking app, we built a foundation for financial literacy that grows with every child.
Built With
- api
- firebase
- firestore
- html5
- node.js
- ollama
- react
- tailwindcss
- typescript
- zod

Log in or sign up for Devpost to join the conversation.