EchoVault

Inspiration

Most people learn about financial history in textbooks — dry facts about crashes and crises that feel distant and irrelevant. But what if those same events could personally warn you about your own financial blind spots? We were inspired by the WEHack theme "Night at the Museum" to bring financial history to life — not as a lesson from the past, but as a mirror for your present.

What it does

EchoVault is an AI-powered personal finance app that uses historical financial crises to teach users about their own money habits. When you sign up, our AI assigns you a Historical Twin — a real type of person from history whose financial situation matches yours today. You receive a Recession Proof Score, get warned by our Déjà Vu Engine when your spending mirrors dangerous historical patterns, and can simulate how you would have survived the 2008 housing crash. Our AI Coach then gives you brutally honest, historically grounded financial advice personalized to your profile.

How we built it

We built EchoVault using React Native with Expo for the mobile frontend, and integrated the Claude AI API for all intelligent features. Our AI pipeline includes five core functions: Historical Twin assignment, Recession Proof Score generation, Déjà Vu pattern detection, Crash Simulator outcome modeling, and an AI Coach chat with streaming responses. We used Supabase for user data persistence and the Wikipedia API for our Daily Crisis Card feature.

Challenges we ran into

Integrating AI responses into a mobile UI in real time was our biggest challenge — ensuring the Claude API returned consistent JSON that our frontend could reliably parse took significant prompt engineering. We also faced API access issues during development due to network restrictions, which forced us to rapidly pivot between providers before landing on a stable solution. Coordinating a 4-person team across parallel branches without breaking each other's code required strict Git discipline.

Accomplishments that we're proud of

We're most proud of the AI Coach's responses — when we asked it about investing in crypto, it responded: "You're about to do exactly what factory workers did in 1929 right before the crash." That moment told us we had built something that genuinely feels intelligent and historically grounded, not just a generic chatbot. We're also proud of shipping a fully integrated 4-screen mobile app in under 24 hours.

What we learned

We learned that prompt engineering is as important as the code itself — getting the AI to return consistent, structured JSON required careful iteration. We also learned that in a hackathon, pre-caching your best AI responses before the demo is not optional — it's survival.

What's next for EchoVault

We want to expand the Crash Simulator to cover more historical crises — the 1929 Great Depression, the 1970s stagflation, and the dot-com bubble. We also want to add real bank account integration so the Déjà Vu Engine can analyze actual transaction data instead of self-reported numbers. Long term, we believe EchoVault could be a genuinely impactful financial literacy tool for first-generation college students and underserved communities who never received financial education.