Inspiration
Our inspiration for creating UW Scavenger came from our own experience at Hack the North 2024. Navigating the sprawling E7 building was overwhelming, and finding quiet, comfortable spots for work and rest felt like an endless challenge. That's why we developed UW Scavenger - to transform the daunting task of exploring campus into an engaging adventure.
With UW Scavenger, you and your friends can embark on three exciting scavenger hunts (ALL WITH AI-GENERATED RIDDLES AS CLUES!!) that will guide you to the best places on campus. Whether you’re searching for a quiet spot to focus or a vibrant place to unwind, our app makes discovering these spots both enjoyable and competitive. Plus, you can connect with other teams and new people who are also on the hunt. Just invite them to join the fun and let the scavenger hunts lead you to the perfect spots for work and play.
By turning campus exploration into a game, UW Scavenger ensures that finding the ideal location is not just easier but also a lot more fun!
What it does
UW Scavenger transforms your campus experience into an engaging adventure by offering three dynamic scavenger hunts. Each hunt is designed to lead you to unique, hidden, and essential spots around campus. Here’s how it works:
Discover Hidden Gems: Our app guides you to the best spots on campus—whether you're looking for a quiet corner to study, a cozy place to relax, or a lively area to socialize. Each hunt provides clues and challenges that reveal these hidden gems in a fun and interactive way.
Compete and Collaborate: Join the scavenger hunts solo or with friends, and compete against other teams for the best times and highest scores. You can also team up with new people you meet along the way, making it a great way to connect and collaborate.
Engage and Explore: With each hunt, you’ll uncover interesting facts, adding an educational twist to the adventure. The app also encourages exploration by leading you to less-traveled areas that you might not have discovered otherwise.
AI-Generated Riddles: We use Groq's API to generate unique and engaging riddles for each scavenger hunt. These AI-created clues add an element of mystery and challenge, ensuring a fresh experience every time.
Interactive Map: Our integration with Mappedin’s HTN 2024 indoor map helps you navigate campus efficiently. The map shows your proximity to your target spots, enhancing your scavenger hunt experience with real-time guidance.
Track Progress and Achievements: Keep track of your achievements and see how you rank against other teams. The app features leaderboards and badges to celebrate your progress and motivate you to continue exploring.
UW Scavenger turns the task of navigating campus into an enjoyable game, making your experience more engaging and memorable. A host chooses three items in three different rooms in the building. They then assign (by sticking a note or writing on/near the item) "item codes", which are numbers used to verify that players found the correct item based on the AI-generated clue provided by Groq's API.
How we built it
We used a Flask server to host the backend, and HTML/CSS/JS (with Bootstrap). We made use of Mappedin's HTN 2024 indoor map to show the players how far away they were from the room they needed to be in. We also used Groq's API, so that a description of the item could be turned into an AI-generated riddle to be solved as a clue by the players. We had no database, and all user data was stored in a dictionary within the server (and in user cache).
Challenges we ran into
We were unable to implement Genesys's API for a feedback system, since the documentation was difficult to interpret and mentors were unavailable, so we were forced to drop that idea and focus on bettering our AI riddles and map technology.
Accomplishments that we're proud of
We are proud of the fact that we were able to adopt a very innovative and subtle use of AI in our project. Instead of making a loud and fancy chatbot, we generated concise (yet an enjoyable exercise for the mind) riddles using Groq's API. This was helpful because it meant the host could choose ANYTHING they would like to be the item, and users didn't see the same boring riddle/clue every time.
What we learned
We learned how to implement Groq's API. This was exceptionally helpful for us because we found that Groq allows us to train the system to behave a specific way for the user. This makes it very easy for developers to launch LLM-based applications, instead of spending a large amount of time fine-tuning an LLM.
What's next for UW Scavenger
We hope that hackers continue to use our idea and benefit from the concept in future Hack the North events. This product has benefited us greatly during HTN 2024 - and we're confident this (and similar applications) could benefit many others. Outside of just HTN, Scavenger could also be used to introduce students to new campuses and ensure that they know important locations such as help desks or fire exits :)
Log in or sign up for Devpost to join the conversation.