Sherpa

Inspiration

The inspiration behind the creation of Sherpa stemmed from a personal experience within our team. One of our members went on a family vacation and found themselves at a loss when it came to planning activities during their trip. They were often overwhelmed by the sheer abundance of options or just genuinely did not know what activities were available at their destination. It was this very challenge that motivated us to develop Sherpa, a solution designed to ensure that such problems no longer hinder your travel experiences.

What it does

Sherpa is an app designed to simplify your travel experience. By simply inputting the number of days and your destination city, Sherpa crafts a curated itinerary for your trip. Say goodbye to the uncertainty of planning activities; Sherpa ensures that you'll always have an exciting list of options at your fingertips.

How we built it

The frontend is built with SwiftUI and utilizes MapKit, the Calendar component, and many other native Apple components. The backend leverages cloud deployed large language models to generate dynamic itineraries for over 4,000 cities.

Challenges we ran into

Our primary challenge revolved around integrating OpenAI's API into the Swift programming language. It was a task that required a blend of prior Swift expertise from some team members while others had to quickly get up to speed. We were operating with Xcode, however, this presented an obstacle, as only a portion of our team had access to Macbooks. Xcode's exclusivity to Apple hardware posed certain limitations for our project, especially for those team members on Windows-based systems. Another challenge was manipulating GPT-3.5's responses to output a consistent response to our prompt.

Accomplishments that we're proud of

We are proud that we completed something in under 24 hours with the challenges we ran into with the limited skillset we had entering this competition. We are also proud that we followed through with the project and committed something even though there were times we wanted to give up.

What we learned

Learning Swift
Integrating OpenAI's API into Swift
Using OpenAI's language model to return a response in a specific way we want

What's next for Sherpa

We would like to implement a way for user's to discard an activity they don't want and replace it with something else. It would be able to gather the user's input and cater their response based on what kind of activities the user discarded. We would also like to input user's feedback in order to improve our app.