• 

  Set up a family account to begin tracking expenses and managing budgets.

• 

  Review family purchase requests and continue through the configured PayPal Sandbox payment flow.

• 

  Access your family’s expense records, budgets, and transaction statuses.

• 

  Review family spending, budget activity, and recent transactions in one place

• 

  Record expenses and review transaction statuses, including refunds when verified events are available.

• 

  Create plain-language spending rules to guide family purchase decisions.

• 

  Test draft spending rules against sample transaction details and inspect the evaluation.

• 

  An overview of PennyPilot AI’s tools for family budgets, spending oversight, purchase requests, and financial guardrails.

Inspiration

Families need to balance children’s independence with responsible spending. Many tools show spending totals but offer little context about purchase decisions or refunds. We built PennyPilot AI to help families track expenses, set category budgets, and make spending decisions together.

What it does

PennyPilot AI is a family-finance prototype for parents managing household expenses and children’s spending.

Implemented in the prototype:

• Create an account and sign in.
• Record family expenses and view transaction history.
• Create category budgets and track spending against them.
• View family, budget, and transaction data on the dashboard.
• Review transaction and refund statuses when available.
• Use purchase-request and spending-policy workflows.

PayPal and AI integration:

• PayPal Sandbox supports account-linking and payment workflows. Webhook routes handle transaction and refund events when configured.
• Gemini can support AI-based checks; rule-based logic is available as a fallback.
• PayPal Sandbox flows require credentials and webhook setup. They have not yet been verified as a complete public demo.

How we built it

• Frontend: Next.js for account, dashboard, family, budget, transaction, and request pages.
• Backend: FastAPI and SQLAlchemy for authentication, family-scoped records, expenses, budgets, and transaction workflows.
• Payments: PayPal Sandbox APIs and webhook handlers for payment and refund lifecycle events.
• AI: Gemini-supported checks, with rule-based fallback behavior.

Challenges we ran into

• Protecting financial records with authentication and family-level access controls.
• Keeping expense and budget totals consistent as transactions are recorded.
• Handling refunds as part of a transaction’s lifecycle, rather than treating purchases as one-time events.
• Using AI carefully: financial decisions need clear rules and human review, not unexplained model output.
• Defining a realistic prototype scope without presenting unfinished integrations as complete.

Accomplishments that we're proud of

• Built an end-to-end prototype for recording expenses and tracking category budgets.
• Connected dashboard and transaction views to backend family data.
• Added PayPal Sandbox integration points for payment and refund workflows.
• Designed the product around both spending visibility and family decision-making.

What we learned

• Financial products depend on trust, clear status information, and secure access.
• Refund visibility matters as much as purchase tracking.
• AI works best as decision support alongside understandable rules.
• Clearly communicating prototype limitations is part of building a credible product.

What's next for PennyPilot AI

Planned improvements:

• Configure and test PayPal Sandbox checkout, webhook verification, and refund flows end to end.
• Add complete family invitations, email verification, and parental-consent flows.
• Improve budget trends, forecasting, and spending insights.
• Refine AI recommendations and make their reasoning easier to understand.
• Deploy a stable public demo with persistent storage.

No additional sponsor tools are currently integrated.

Built With

• budgeting
• fastapi
• fintech
• gemini
• javascript
• machine-learning
• next.js
• oauth
• paypal
• paypalapi
• paypalsandbox
• postgresql
• python
• refundtracking
• sqlalchemy
• typescript
• webhooks