UniMatch

Inspiration

In 2025, a startling study made headlines: almost two-thirds of college students are lonely. As freshmen at Princeton University, we have heard the same thing over and over again: I feel alone. Whether it’s the hopeless singles who long for a date on Fizz (an anonymous campus conversation site) or lonely students who lament about spending Friday nights in their dorm, hundreds of students long to find their “match.” We’ve felt this firsthand as well. Countless social mishaps, from standing awkwardly at eating clubs to fumbling a date after being unsure of what to do, have left us wanting more. We wanted to encourage social interaction in a novel way. Enter UniMatch. UniMatch is built for students, by students, ensuring that it is receptive to our needs. Derived from UNIversity MATCHmaking, we aim to excite campus culture by facilitating human interactions.

What it does

UniMatch is not your average dating app. You don’t have to spend hours swiping left or right. Instead, users follow a magic link to their homepage, where they fill out a 20-question survey, similar to the BuzzFeed or UQuiz surveys that many of us fill out as a hobby. On the home page, users can choose between two distinct surveys. One can match you with a romantic interest, while the other can match you with a platonic friend. This lets users pick what type of companionship they seek. The quiz’s questions have been curated to have playful language while also revealing personality traits, communication styles, and values, so that the user can enjoy filling out the survey while also sharing the information we need to match you. The survey can be quickly completed since it uses a multiple-choice answering system that allows you to rate statements from strongly agree to strongly disagree. We modeled this system off the 16-Personalities Quiz due to its success in having users complete it. The algorithm uses a weighted sum to compare answers between users and finds a one-to-one match based on similar compatibility scores. On Monday evenings at 8 pm, the administrators (us) will hit the matchmaking button on the admin site to push out users’ matches. This increases anticipation for the event. Apart from receiving their match’s name and email, they will get a set of AI-generated date ideas based on their shared interests. This solves another problem we faced as college students: figuring out how to plan a date. By giving them curated suggestions that are unique to the area, students can have a stress-free time hanging out with their match. Additionally, if a user finds their soulmate through the site, they have the option to opt out of the next Match Monday on the homepage. Apart from the matchmaking aspect, our site facilitates entertainment on the site, not just off of it, by gameifying social connections. While traditional social networking sites have people competing against each other for likes, our system rewards users who celebrate connections with their friends. Users can “ship” their friends, and if the algorithm ends up pairing them together, the “captain” of this ship will receive points. These points could land them on a Top Ten Captains leaderboard, which displays the points and names of the users with the most points. These Captains will also receive a prize at the start of each month. This “shipping” feature has users coming back for more and also makes the idea of matchmaking a little more magical. Also, since people will want to ship their friends, the site can grow organically through word-of-mouth. The final feature is a secret admirer system. Many of us are too scared to approach our crushes in real life. The “Fishing Dock” allows users to anonymously send their crushes an email letting them know that “someone” has a crush on them. This helps build courage and encourages initiative in students, allowing them to forge the relationships they desire.

How we built it

With Claude. ai as our primary coding source, UniMatch is built with Next.js and React for the frontend, which provides a fast responsive user experience with smooth animations and intuitive navigation. This stack gives the ability to create interactive interfaces that college students expect from modern web applications. For our backend and database, we used Supabase, which gives us real time database capabilities, authentication, and easy to use API’s. The database was structured with multiple tables: -responses table to store user survey answers, preferences, and demographic information -matches table to store matched pairs with compatibility scores and AI-generated date ideas -ships table to track which users shipped which friends together -captains_leaderboard table to maintain the gamification leaderboard with points and stats -secret_admirers table for our anonymous fishing feature where users can shoot their shot Authentication was implemented using Supabase’s magic link system, which allows a passwordless, frictionless login experience that is secure and easy for students to use. Users simply have to enter their Princeton email and will receive a one time login link. The matching algorithm works by calculating compatibility scores through a weighted point system:

Base compatibility(up to 20 points for dating, 19 for friends) comparing survey answers where matching responses earn full points and close answers earn partial points.
Hobby bonus (up to 5 points) rewards shared interests from the user’s selected hobbies
Shipping bonus (up to 5 points) added extra weight when captains have “shipped” two people together, with diminishing returns for multiple ships The algorithm also takes into account user preferences, like filtering matches by gender preference, age range, and height preferences for dating matches. For the AI-powered date suggestions, we integrated Groq’s LLM API using the Llama 3.3 70B model. When matches are created, the system looks at both user’s shared hobbies and personality alignments from the survey responses to generate four personalized Princeton specific date ideas. The AI considers factors like budget(college friendly), location(campus), and the type of relationship(romantic or platonic). The admin dashboard lets us run the matching on demand, view submissions, and see the results of matches with compatibility percentages and AI generated suggestions. This lets us monitor the platform’s success on “Match Monday” and gives us full control to make sure everything runs smoothly. We CSS animations, specifically heart shaped boat animations that float across the shipping and leaderboard pages, which makes the “captain” theme more pronounced and makes the user experience more entertaining. The development process followed this order:
Survey system - Building the 20 question personality quizzes and other preference questions
Database schema design - Planning all tables and relationships in Supabase
Authentication through implementing magic link login
Matching algorithm - creating the compatibility calculation
AI Integration - Connecting Groq API for personalized date suggestions
Admin dashboard - building the interface for running the matches
Results page - Shows the matches to users
Shipping/Captain system - Added gamification with leaderboard
Fishing feature - Implements anonymous secret admirer system
UI polish - adding animations and making the UI look cleaner.

What we learned

As first-time hackers, this experience was extremely illuminating. We learned to effectively troubleshoot, utilize LLMs to code an application, understand high-level Javascript code, and explore web design.

What's next for UniMatch

The first order of business would be releasing UniMatch for Princeton students to use through the ListServ, Fizz, and Instagram. This way, we can start achieving the impact we want. While UniMatch is currently tuned to Princeton students, we hope to turn this website into an IOS app that uses students’ email addresses to connect them to a version of UniMatch that is specific to their campus. This way, every college student across the country and even the world can reap the benefits of our idea. Not only that, but we would like flagship UniMatch administrators at each of these universities so that they can reach out to local businesses and restaurants for sponsorships. These sponsorships would ask for donations from these companies in exchange for a boost in visibility in the algorithm when it comes to the date recommendations. This symbiotic relationship would allow us to monetize our platform with paywalls, which would allow users to access every feature of the app for free. Since many big tech dating apps have deteriorated in quality due to their paywalls, we hope to stand out as one that won’t sacrifice real connections for profit. Finally, we hope to integrate AI more into the app. After curating a larger database, we hope to train a machine learning model to match users and also integrate an AI chatbot for live advice before, during, and after the date. We want users to feel supported, not just in the matchmaking process but throughout their relationship. Sustaining a relationship can be just as hard as starting one, and we want to help users in every step of the process.