BoilerBets

Home Page
Rotating Live Bets on home page
Dashboard of bets
AI chatbot to help with bets
Betting page
Hamburger dropdown
Sign In page
Leaderboard of most coins

Overview BoilerBets is a fun, campus-focused prediction game designed to feel uniquely Purdue. It lets students place fake-currency bets on campus events like sports, arts, and competitions. The goal is to help students learn basic odds thinking while engaging more with school life—without using real money.

What It Does BoilerBets is a Purdue prediction platform where users can create events with multiple outcomes, bet fake coins before deadlines, earn payouts through pool-based distribution when events are resolved, and climb a leaderboard based on total balance. Users can also ask Pete, an AI chatbot powered by Gemini, for betting advice and strategy. The platform runs entirely on fake currency.

How We Built It The frontend uses Next.js 16 (App Router), React 19, Tailwind CSS, and shadcn-style components. The backend uses Next.js API routes, MongoDB Atlas, and Mongoose. Authentication is handled through Auth.js (NextAuth v5) with Google sign-in. Gemini powers the Pete betting advisor, and the app is hosted on Vercel. The sports events are also automatically updated through the ESPN API.

Challenges We Faced We dealt with Gemini model availability issues by switching to gemini-2.5-flash. Pete initially sounded too generic, so we refined system prompts to produce more varied, contextual responses. We also fixed undefined balance errors after database test data deletions by adding null and undefined checks.

Accomplishments We built a full prediction platform with authentication, events, betting, payouts, and AI integration. We successfully implemented pool-based payouts with transaction safety and created a Purdue-focused design and user experience.

What We Learned We learned how to use the Gemini API and shape AI behavior with prompt engineering, implement pool-based payout math with atomic MongoDB transactions, handle missing data in real-time UIs, and fully utilize the Next.js App Router with server components and API routes.

What’s Next Next steps include email verification for new signups, more event categories and filters, push notifications for event results, improved Pete AI with event-specific context and memory, and automatic syncing of Purdue sports data from external APIs.