Inspiration

Events, such as hackathons, require a lot of planning. One part of planning an event is getting all the resources you need. The research and management of resources is typically done by hand, and is a very effort and time consuming process. So, we unified the tools necessary and automated this process using AI.

What it does

EventPrep takes a description of your event, and suggests resources that you need. Then, it searches the internet for difference sources and allows you to select which sources you are going to buy from. The TodoList allows you to keep track of which resources you have bought, chosen to use, or are still considering.

How we built it

We used Next.js to host the website, the OpenAI API with the web search tool to allow for internet research, and used Supabase for both authentication and database.

  1. The database is the source of truth, and the frontend pulls from the database. However, a separate layer of data is kept on the frontend to allow instant reactions to user interaction, while saving that data is set on a delay.
  2. We used the web search tool and a json schema to get reliable data from the OpenAI API, which we then processed and added to the database.
  3. Supabase subscriptions were used to keep data fresh and also used to trigger the automatic searching of resources.

Challenges we ran into

This was the first or second hackathon for half our team, so we were still learning a lot of things. Additionally, the more experienced team members decided to get a bit ambitious and tried to tackle things they've never done before. We learned a lot about programming the frontend, using AI APIs, and integrating frontend flows with database flows. We are also all students, so we did not have a lot of time during the week to work on this week long hackathonl

Accomplishments that we're proud of

I am most proud of getting the frontend to work with the database since it was my first time trying to do this. I felt like my architecture designs made it possible to complete this task, and I am most proud of that.

What's next for EventPrep

We can polish the frontend, maybe add some more color. We can also add better indicators of when the AI agents are running. We also want to add rate limiting or some sort of limit to the AI calls.

Built With

Share this project:

Updates