springbreakr

A BostonHacks 2024 project

Introduction

Have you ever struggled with coming up with travel plans? Introducing Spring Breakr, an AI-powered travel assistant! Give it a short description of what you are looking forwward to do, and it will recommend a series of destinations from which you can pick and choose from! First, pick a recommended country which you want to visit, then pick from among a list of destinations specially tailored for you within that country!

How it works

We are using the LLAMA LLM model hosted on a cloudflare worker to interpret user sentiment and recommend destinations. The frontend is built using Next.js, and the frontend and backend are connected through APIs we built in Flask. We also have a travel category prediction AI we trained using KNN, which we intended to use for user data anonymization. With this technique, we do not save user prompts directly in the database, only the category of their travel.

How to run

cd backend pip install -r requirements.txt python app.py cd ../client npm install npm run dev