EzCite

Inspiration

We were inspired by citation sites such as https://www.citationmachine.net/ and https://www.easybib.com/ and wanted to use the power of specialized LLMs to see if we could not only find relevant sources for the user based on their background knowledge, but also adequately save time for the student and cite the source for them.

What it does

EzCite uses artificial intelligence and puppeteer to identify potential citations in the input, and finds potential relevant links and cites them for the user.

How we built it

We built our website front-end using React.js. Our back-end is built using a mix of LLMs, discord.js, and puppeteer.

Challenges we ran into

The main challenge we ran into was not being able to incorporate the algorithm we developed into our website's front end. Puppeteer has a very hard time compiling into Netlify since it is a serverless host. In the end, we made a Discord bot to incorporate our algorithm into.

Accomplishments that we're proud of

We are proud of the fact that we have learned so much in such a short amount of time, and were able to put together a somewhat working product that we all thought was useful.

What we learned

We learned to use React.js to build a good looking website, puppeteer to navigate a website without visual confirmation, and how to use range of APIs.

What's next for EzCite

Our very next step is to implement the backend to work with the front end, by hosting it on a full server. Furthermore, we aim to improve the accuracy of our algorithm by prompt engineering and developing an in-house citation scrapper. Eventually, we hope to mature our product and give everyone an easier time when citing their sources.

Built With

• axios
• citationmachine
• discord.js
• javascript
• node.js
• openai
• perplexityai
• puppeteer
• react