Penny

Inspiration

Penny began with one question from a mentor: “If I gave you $100 to invest, what would you do with it?”

One of us answered, “Nothing.”

That changed the direction of our project. The problem was not a lack of investing information—there are already countless tutorials, definitions, and trading platforms. The real barrier was the fear of making the wrong decision, losing money needed for everyday life, and feeling responsible for choices you do not understand.

We stopped building another educational tool and started building a companion that could safely take on that burden.

What We Built

Penny is an iMessage-first automated investing companion that puts the user’s life before their portfolio.

A user can text:

“I received $100, but I need $60 for textbooks.”

Penny remembers the expense, protects the $60, and offers a simple Low, Medium, or High-risk plan for the remaining $40. After the user confirms the proposal and verifies their identity, Penny sends paper-trading orders to a connected brokerage sandbox.

Penny then maintains the relationship by sending weekly updates and adapting when life changes. A raise could create room to invest more, while a car breakdown could cause Penny to pause contributions, reduce risk, or propose a withdrawal.

Unlike a financial chatbot, Penny does not simply explain what someone could do. It maintains a living financial mandate and carries out approved actions.

How We Built It

Photon’s Spectrum SDK gives Penny its conversational home in iMessage.

Gemini translates natural language into structured facts such as income, expenses, life events, and risk preferences. Gemini never controls the money. A deterministic safety engine calculates the protected and investable amounts, then builds a transparent portfolio from predefined allocations.

Auth0 verifies the person approving consequential actions. PostgreSQL provides long-term memory for each user’s mandate, protected expenses, confirmations, brokerage references, and portfolio history. Alpaca’s Broker API sandbox creates the fictional brokerage relationship and executes virtual-money ETF orders.

Replit hosts the backend that connects these services. We also created an iMessage-style browser prototype as a reliable presentation environment, but the intended product experience lives in Messages.

Challenges We Faced

Our greatest challenge was separating helpful automation from irresponsible automation.

Penny needed to understand natural conversation without allowing an AI model to make uncontrolled financial decisions. We addressed this by limiting Gemini to fact extraction and placing every financial calculation behind deterministic, testable rules.

We also had to design for duplicate messages, expired approvals, identity verification, persistent memory, unexpected life events, and brokerage failures. Every proposal is tied to a specific user, mandate version, confirmation code, expiration time, and verified identity.

Finally, we had to present a realistic financial product without pretending that a hackathon prototype could legally manage real customer money. All funds and trades in our demonstration are clearly labeled as simulated.

What We Learned

We learned that accessibility is not just explaining complicated terminology more clearly. Sometimes accessibility means removing unnecessary decisions entirely.

We also learned that responsible AI systems need clear boundaries. Gemini is valuable because it understands the user’s language, but Penny’s safety comes from deterministic rules, explicit consent, verified identity, persistent records, and a regulated brokerage boundary.

Most importantly, we learned that an investing product should begin by asking what someone needs their money for—not which stock they want to buy.

What’s Next

Our next step is to unify the complete Photon, Auth0, PostgreSQL, and Alpaca sandbox experience into one polished iMessage flow.

Beyond the hackathon, launching Penny with real money would require licensed advisory and brokerage partners, full KYC/AML onboarding, custody, disclosures, regulatory supervision, security review, and additional user protections.

Our long-term vision is an investing companion that grows with the user: remembering their responsibilities, responding to changes in their life, and helping them build wealth without requiring them to become a financial expert first.

Most investing apps start with the market. Penny starts with your life.