bookfinder

generated by gemini

Inspiration

Bookstores offer incredible value by way of ambient discovery. When you have access to the stacks as a customer, you can't be marketed to in the same way as through shopping feeds, where SEO and marketing can entirely cover up the existence of entire categories of products. I wanted a way to achieve a curious discovery mechanism with tech.

What it does

Bookfinder takes a single word subject query, then retrieves suggestions by a set of familiar categories:

well regarded academic publishers
common types of genre fiction
childrens books
newest published books

But then, we use Gemini to identify and retrieve related sets of books, unsupervised, and visualize those arbitrary categories to add an element of curiosity!

How we built it

I chose to use React because it's a rich ecosystem for creating webapp experiences. Then, I asked gemini to write me a lot of code, repeatedly...

Challenges we ran into

I wasted a huge amount of time attempting to write this project in Vite, because Gemini said that was the new, best build system for React apps. I eventually became convinced Vite is not designed for full-stack applications, where the app endpoints that hit private APIs are also written in JS, and the mentors confirmed that. So I moved to Next.js