TappClub

Inspiration

Going out with friends is super fun, but actually coordinating an event can be super annoying. Often, everyone in a friend group might be free, but none of them feels like taking the time to schedule something. The friction to event-planning is immense: social media apps focus on maintaining their users by emphasizing doom scrolling, but our focus is to facilitate meaningful outings for our users.

What it does

Our goal is to optimize the before, during, and after of an event experience. We believe in changing the way people use social media: encouraging them to stop doom scrolling and go touch grass. That's why TappClub only allows users to post by actually attending events, and tapping to confirm their attendance. Through TappClub, users can also plan and organize their events, all with agentic assistance in one seamless experience.

How we built it

We used React Native with Expo to build a cross-platform mobile app, because we all had experience with web technologies. We used Postgres with Neon for our database, HCE (Host-based card emulation) and NFC (Near Field Communication) tags with the ISO-DEP protocol for tap detection, Auth0 for authentication, and Gemini with search-based grounding for the agentic workflow. We used Python with FastAPI for the backend, hosted on Render, with Websockets for real-time notifications.

Challenges we ran into

Each technically-focused member on our team had their "major problem" to tackle. Davis tackled NFC reading on the front end, which was especially difficult with the cross-platform React Native, which needed custom Expo plugins to implement NFC reading and HCE, and implemented functions following the ISO-DEP protocol. Sragvee handled secure authentication, which involved securing the entire application and linking Auth0 to our custom Neon database. Arnav researched and implemented a full Gemini-powered event planning workflow, which needed custom context from grounding and user calendars.

Accomplishments that we're proud of

We are especially proud of making extremely fast progress on simple "boilerplate" code, which took relatively little time to complete. We each discovered the effectiveness of various AI tools such as Cursor, Copilot, and Roo, and were able to spend the majority of our time tackling the difficult, interesting problems. Specifically, Sragvee was able to leverage Cursor with a detailed "specification" prompt, which allowed us to scaffold the entire app in less than an hour and move on to our independent tasks, avoiding merge conflicts.

What we learned

We all learned how to use generative AI code editors to heavily speed up our workflow. Additionally, we improved our React Native, Python, Neon, Gemini, and time management skills. Collaborating within such a tight timeframe taught us how to coordinate extremely well while maximizing the productivity of every hour of our time.

What's next for TappClub

In the future, we plan to improve performance by moving a lot more processing over to the server. We also aim to collect more sources of data in order to better recommend events to our users, and potentially monetize via affiliate programs with partnered events. Finally, we will publish the app to both the Android and Apple app stores and purchase a paid Apple Development Account in order to integrate NFC functionality on iPhones.