• 

  Architecture Diagram

BookBuddy - AI Reading Companion

Inspiration

I was trying to find a good book to read but got overwhelmed by all the choices on Amazon. I just wanted someone to tell me "read this book, it's perfect for you." That's when I thought - what if I could build an AI that does exactly that using AWS Bedrock?

What it does

You ask BookBuddy for book recommendations like "I need motivational books" and it gives you specific titles, authors, descriptions, and Amazon links to buy them. It's live on the web right now - anyone can try it.

How we built it

Built with AWS Bedrock Agents using Claude 3 Haiku, Python, and Streamlit for the web interface. The AI gives book recommendations, then my code cleans up the responses and automatically generates Amazon purchase links. Deployed on Streamlit Cloud so it's accessible to everyone.

Challenges we ran into

Learning AWS Bedrock Agents from scratch was tough. The AI responses were messy and inconsistent, so I had to write code to clean them up. Getting Amazon links to work with weird book titles was tricky. Also forgot to enable model access in AWS console at first!

Accomplishments that we're proud of

It's live and working! Fast responses (under 2 seconds), reliable Amazon links, and a clean web interface that works on mobile. The code is organized and has good error handling.

What we learned

First time using AWS Bedrock Agents - learned how to set them up and handle messy AI responses. Also learned the difference between building something for myself vs. building for other people to actually use.

What's next for BookBuddy

Add user accounts to save favorites and personalize recommendations. Include more bookstores beyond Amazon. Better summaries with plot details. Maybe a mobile app and social features for sharing recommendations with friends.

Built With

• bedrock
• python
• streamlit