The Lending Shelf

A peer-to-peer, decentralised library which gives dusty bookshelves and finished books a new life, encouraging community connection and cultivating excitement and passion for reading.

Comment

The problem we were inspired to solve

Information has become more accessible than ever. Yet, our society hasn't become more informed.

  • Ebooks offer us information almost instantly, with just a couple of clicks.
  • Council owned libraries offer extensive reading resources.
  • Most households own books.

But have we become more well read? Certainly not. The habit of reading for leasure has fallen out of fashion. Only ~11.5% of young Australians read on any given day (ABS, 2024)

Our proposed solution

If the increase of information didn't help reading flourish, then what is the way forward? We think it could be a free peer-to-peer book exchange platform. While our solution may lack the extensive corpus and range that traditional libraries and ebooks provide, it aims to make up for it by cultivating excitement, passion, and social connection.

How we built it

  • We built a web app.
  • The front end was created with React.
  • The back end was created with Node.js/Express, Typescript, PostgreSQL hosted on an AWS instance, and hosted with - nginx.
  • We used Gemini Vision Model API (Flash-2.5) to automatically extract and upload book metadata from images of bookshelves. Preston Hinkle's dataset helped us during the development and testing of this feature.
  • We used OpenLibrary API to extract book metadata.

Challenges we faced

  • We struggled with ideation. We took the entire first night to finalise our idea. It took a lot of effort to discuss and evaluate ideas without bias. We often found ourselves back at square one after some preliminerary research found that an idea was not as novel or useful as we first thought.
  • We were challenged by our poor initial choices of technologies/libraries, as this lost us time. For example, we wrote our backend with basic libraries like pg, which required manually writing SQL schemas/queries and auth, as opposed to using a tool like Supabase which automatically handles a lot of the boilerplate.

Built With

Share this project:

Updates