Platform Image
Clusters Layout
Main MapBox indicating all the clusters
Leaderboard based on Amount of money saved
Manual Fare Split Calculator
Campus Green Impact of Car Pooling
Upcoming Rides Pages
Manage Rides Page
AI Chatbot

WolfiePool

The Paradigm Shift in AI-Powered Campus Mobility for Stony Brook University

Forged with Gemini API · ElevenLabs TTS · Supabase · Next.js · FastAPI · Auth0

📖 The Epic Behind WolfiePool

💡 The Genesis: A Broken Commute

The modern collegiate commute is fundamentally flawed. Every single day, hundreds of Stony Brook University students bleed their limited funds on surge-priced, single-occupancy rideshares. Whether commuting from the LIRR train station, returning from a late-night grocery run, or heading home from the library at 2 AM, students are unwittingly participating in a massive, uncoordinated logistical failure. They travel in the exact same directions, at the exact same times, yet pay individual premium fares while collectively pumping unnecessary carbon emissions into the atmosphere.

Group chats and Discord servers attempting to coordinate shared rides are chaotic, unreliable, and often compromise safety. We realized that solving this didn't just require an app—it required a comprehensive, mathematically rigorous, and AI-driven ecosystem. We needed to replace the chaos with intelligent orchestration. Thus, WolfiePool was born.

🚀 The Vision: What It Does

WolfiePool is not merely a ride-sharing app; it is a hyper-localized, intelligent transit orchestration platform designed exclusively for the Stony Brook University community.

Students input their destination and preferred time window. From there, the platform takes over. Our bespoke geospatial clustering algorithm evaluates bearing angles and Haversine distances in real-time to group up to four compatible riders heading in identical trajectories.

But true innovation lies in equity. We completely discarded the archaic "split the bill evenly" model, which unfairly penalizes riders with shorter trips. Instead, WolfiePool dynamically calculates distance-weighted fare splits, ensuring every cent paid is strictly proportional to the rider's specific drop-off distance.

To elevate this from a utility to an experience, we injected cutting-edge AI and Voice technologies:

Gemini 2.5 Flash serves as the cognitive engine of the platform, powering "Ask Wolfie" (a context-aware global campus assistant), predictive ride-matching analytics, and personalized environmental impact reports.
ElevenLabs API transforms the platform into an immersive, hands-free experience. High-fidelity Text-to-Speech (TTS) natural voice notifications alert users of ride matches and deliver unignorable, context-aware safety alerts during late-night hours.
Supabase Realtime synchronizes the entire platform instantaneously, providing matched riders with immediate, secure in-cluster chat to coordinate their precise pickup logistics.

🛡️ Uncompromising Security & Trust

A campus platform is only as viable as its safety model. To guarantee an impregnable wall of trust and exclusivity, we engineered a zero-compromise authentication pipeline. Utilizing Auth0, we established a strict gatekeeping mechanism: Only verified Stony Brook University organizational accounts can authenticate. Furthermore, this is tightly coupled with Duo Security SSO, ensuring that every user on the platform is a multi-factor authenticated, legitimate member of the SBU community. No outsiders. No fake accounts. Pure trust.

🛠️ The Engine Room: How We Built It

We architected WolfiePool to be a masterclass in modern full-stack development, prioritizing absolute scalability and flawless user experience:

The Frontend: Built on the bleeding edge of web technologies utilizing Next.js 15, React 19, and Tailwind CSS v4. We crafted a deeply responsive, mobile-first interface featuring native dark-mode and visually stunning, interactive Mapbox GL routing arcs.
The Backend: Driven by the high-performance async capabilities of FastAPI (Python 3.11+). This server operates as the mathematical core, executing our complex bearing-angle logic (mandating a ≤30° trajectory spread and ≤2.5× distance ratio) while serving rapid REST endpoints.
The Database: Supabase (PostgreSQL) acts as our robust data layer, utilizing row-level security and WebSockets for real-time state mutations and live messaging.
The Integrations: A seamless symphony of REST API calls bridging Google's Generative AI, ElevenLabs' audio generation, and Mapbox's geospatial rendering.

⚠️ The Crucible: Challenges We Conquered

Building a platform of this magnitude within the constraints of a hackathon tested our engineering limits:

The Geospatial Clustering Dilemma: Traditional distance proximity is useless for ride-sharing. Two students might be 5 miles away from campus, but if one is heading North and the other South, clustering them is disastrous. We had to dive deep into vector mathematics, implementing spherical trigonometry to calculate bearing angles, ensuring riders were actually moving in the same geographical corridor before calculating distance ratios.
The Equity Equation: Formulating a dynamic pricing algorithm that felt definitively "fair" required meticulous tuning. We layered a proportional distance-weighted mathematical model over simulated Uber fare base rates to guarantee exact, unarguable financial savings for every unique route.
Asynchronous Audio Rendering: Seamlessly triggering ElevenLabs' TTS on the Next.js frontend without blocking the main UI thread—especially when dealing with dynamic, AI-generated route summaries and real-time departure countdowns—demanded highly complex asynchronous state management and audio buffering.

🏆 The Triumphs: What We Are Most Proud Of

A Flawless Live Deployment: We didn't just build a local demo; we shipped a production-ready application. Experience it live here.
Symphonic AI/Voice Fusion: We moved beyond gimmickry. Gemini isn't just a chatbot; it's an analytical engine predicting match probabilities and parsing carbon footprints. ElevenLabs isn't just a voice; it's a vital accessibility and safety tool that demands attention for late-night transit alerts.
Tangible Ecological Impact: By actively calculating and displaying actual lbs of CO2 avoided per matched ride, we built a gamified system that tangibly rewards the SBU community for sustainable behavior.
Ironclad Authentication: Successfully implementing the strict Auth0 + Duo SSO flow to ensure our application remains a completely walled, hyper-secure garden for SBU students.

📚 The Epiphany: What We Learned

The sheer power and complexity of orchestrating real-time, bi-directional state synchronization using Supabase WebSockets.
Advanced prompt engineering and context-window optimization to force Gemini 2.5 Flash to adopt the precise persona of a Stony Brook transit savant.
Translating complex geospatial mathematics from Python backend scripts into fluid, beautiful, user-facing Mapbox visualizations.

🔮 The Horizon: What's Next

WolfiePool is just the beginning. Our roadmap for the future includes:

Direct Rideshare API Integration: Bypassing manual Uber booking entirely by directly calling rideshare APIs to dispatch vehicles the moment a WolfiePool cluster reaches capacity.
The "CampusPool" Abstraction: Refactoring our hardcoded SBU coordinates and Duo SSO rules into an agnostic SaaS configuration, allowing any university on earth to deploy their own localized instance of the platform in minutes.
Native Mobile Ecosystem: Translating our responsive web app into fully compiled iOS and Android applications utilizing React Native, enabling deep OS-level push notifications and background location tracking.

🎯 Hackathon Track Highlights

🤖 Best Use of AI / Gemini API

Ask Wolfie – Global AI Assistant: A persistent, contextually aware floating chatbot powered by Gemini 2.5 Flash that processes questions about fares, platform logic, and campus scheduling in real-time.
AI Route Narratives: Gemini dynamically parses matched clusters to generate conversational, human-like trip summaries (e.g., "You'll swing by Dhruv then head straight to the SAC in about 15 minutes!").
Intelligent Safety & Prediction Analytics: Context-aware safety tips generated based on time-of-day and travel distance, coupled with a predictive matching engine that estimates the probability of finding a cluster right now.

🔊 Best Use of Voice / ElevenLabs API

Dynamic Match Announcements: When a student is matched, a hyper-realistic TTS announcement orchestrates the victory: "Great news Vraj! You've been matched with 2 other riders. You'll save $4.50!"
Unignorable Safety Voices: For high-risk, late-night rides (8 PM–6 AM), AI-generated safety tips auto-play via TTS—ensuring critical information is heard, not just skimmed.
Departure Countdowns: A spoken alert fires precisely when a cluster's departure window is 5 minutes away, keeping the entire group punctual.

🌍 Best Sustainability / Social Impact Hack

Verifiable CO2 Tracking: Every shared ride computes the precise estimated carbon savings (lbs CO2 avoided), with a strict logic gate ensuring solo riders correctly show $0 savings.
Financial Equity: Our distance-weighted fare splitting democratizes the cost of commuting, ensuring fairness over equality.
Trust-Based Community Building: Driven by SBU-only Duo Auth0 SSO and Supabase Realtime chat, WolfiePool actively fosters a secure, communicative campus ecosystem.