XPense

XPense turns personal finance into an interactive game, helping young users understand spending, practice investing, build healthier habits, and make smarter financial decisions.

Comment

XPense

We built a system that turns personal finance into a game, helping young users learn, practice, and build healthier money habits through XP, simulation, and interactive feedback.

🧠 Inspiration

Learning finance is important.

But for most young people,

it does not feel fun.

Budgeting feels like homework.

Investing feels intimidating.

Financial apps feel like dashboards made for adults.

But games are different.

In games, we learn by doing.

We earn XP.

We unlock rewards.

We build characters.

We improve because progress feels visible.

So we asked:

Why can't learning money work the same way?

What if saving, budgeting, investing, and building healthy habits felt like progressing through a game?

Not just reading numbers—

but learning through interaction.

Practicing without real financial risk.

Seeing your progress grow with you.

So we asked a simple question:

What if learning finance felt like leveling up?

What it does

XPense is a gamified financial learning system that turns real financial behavior into interactive lessons, simulations, and progress.

Connect → Learn → Improve

🎮 Financial Learning Engine

  • Connect synthetic financial data through Capital One Nessie
  • Analyze balances, spending, debt, and investment activity
  • Turn financial behavior into XP, health scores, streaks, and unlocks
  • Practice investing safely through paper trading
  • Use Gemini to understand user questions while deterministic engines calculate the financial result

At the core, XPense separates AI reasoning from financial truth:

User Question
      ↓
Gemini Router
      ↓
Validated Financial Snapshot
      ↓
Deterministic Engine
      ↓
Verified Result
      ↓
Gemini Explanation

Where:

  • Gemini = understands intent and explains results
  • Financial Snapshot = trusted user financial data
  • Engine = performs the actual calculation
  • XPense UI = turns the result into an interactive learning experience

This lets users learn finance conversationally without letting AI invent financial numbers.

Accounts Swipe Trade Community Safety
XPense accounts screen XPense Swipe Trade screen XPense community safety map

🔥 Why It’s Different

Most finance apps:

  • Track your spending
  • Show your balance
  • Give you charts and reports

XPense does something different:

It turns personal finance into a game you learn by playing.

Key differentiators:

  • 🎮 Learn by doing — earn XP, build streaks, unlock cosmetics, and evolve your financial alter ego
  • 📈 Practice without risk — learn investing through paper trading before touching real money
  • 🧠 Rewards healthy behavior, not just wealth — financial health matters more than simply making money
  • 🔥 Makes spending patterns visible — Map of Time reveals when and where your money leaks
  • 🤖 AI explains, engines calculate — Gemini makes finance conversational while verified engines handle the actual math
  • 🛡️ Built-in financial safety — Persona-backed community fraud reporting helps distinguish trusted signals from spam

XPense bridges financial education + gaming + AI + real financial behavior.

How We Built It

🔁 System Pipeline

Student → Next.js UI → Supabase Auth → API Routes → Financial Data + AI + Safety Services → Supabase Postgres → Realtime Feedback → User

Architecture in one view

XPense

⚙️ Tech Stack

Frontend

  • Next.js, TypeScript
  • Tailwind CSS
  • React

Backend & Database

  • Supabase Auth
  • PostgreSQL
  • Supabase Realtime
  • Redis

Finance

  • Capital One Nessie
  • Alpaca
  • Financial Modeling Prep

AI

  • Google Gemini
  • ElevenLabs
  • Zod validation + deterministic financial engines

Identity & Trust

  • Persona

Maps & Safety

  • Mapbox

Blockchain

  • Solana Devnet
  • Anchor

🧠 Challenges

1. Making Finance Feel Like a Game

We wanted XPense to feel rewarding without encouraging reckless financial behavior.

We solved this by tying progress to financial health, consistency, diversification, and healthy habits instead of simply rewarding higher returns.

2. Keeping AI Financially Reliable

LLMs are great at conversation, but they should not invent balances or financial calculations.

We separated the system into:

  • Gemini for understanding intent
  • Supabase for trusted financial data
  • Deterministic engines for calculations
  • Gemini again for explaining the verified result

3. Preventing Fraud-Report Spam

Community reporting can easily be manipulated by fake or duplicate accounts.

We combined:

  • Persona human verification
  • reporter reputation
  • weighted confidence
  • time decay

This allows trusted reports to matter more than large amounts of low-confidence spam.

4. Making Undoable Trades Reliable

A 10-second undo sounds simple until users refresh, retry, or send duplicate requests.

We moved trade execution into Postgres-backed scheduled actions with idempotency, so each paper trade either safely reverts or commits exactly once.

🎉 Accomplishments

  • Built a full gamified financial learning platform during HackRICE
  • Integrated banking, AI, paper trading, identity verification, maps, voice, and blockchain into one connected system
  • Built a 10-second reversible paper-trading engine with database-backed execution
  • Created Map of Time, turning spending history into a 7 × 24 behavioral heatmap
  • Built an AI advisor where Gemini explains, but verified engines calculate
  • Created an end-to-end experience from financial data → learning → action → progress

📊 Impact

  • Makes financial education feel interactive instead of intimidating
  • Lets young users practice budgeting and investing without risking real money
  • Turns everyday financial behavior into visible progress, XP, and learning
  • Helps users build healthier habits before financial mistakes become expensive
  • Bridges the gap between financial data and real financial understanding

🔮 What’s Next

  • Expand the financial-health engine with deeper real-time behavior tracking
  • Add more investing simulations, challenges, and learning missions
  • Make the XPense alter ego evolve dynamically with financial habits
  • Add smarter predictive insights for spending patterns and risky behavior
  • Expand from sandbox data toward secure real-world financial integrations

🧾 Final Thought

We are not lacking financial data.

We are lacking engagement.

XPense doesn’t just track money.

It turns financial learning into something you can play, practice, and improve at.

And in doing so,

it gives young people something finance rarely gives them early enough:

confidence.

Built With

  • alpaca
  • anchor
  • anchor-other:-zod
  • blockchain
  • css
  • devnet
  • elevenlabs
  • elevenlabs-finance-&-market-data:-capital-one-nessie
  • financial-modeling-prep-identity-&-trust:-persona-maps:-mapbox-blockchain:-solana-devnet
  • mapbox
  • nessie
  • next.js
  • persona
  • postgis
  • postgresql
  • redis
  • redis-ai:-google-gemini
  • solana
  • supabase
  • supabase-realtime
  • tailwind
  • tailwind-css-backend-&-data:-supabase
  • typescript
Share this project:

Updates

Submission history