🚀 Inspiration

Shopping online should be easy, but sifting through hundreds of reviews to find the best product can be time-consuming and frustrating. 🛒⏳

We thought: What if AI could summarize reviews for you? Instead of spending hours reading, you’d get a concise breakdown of the pros, cons, and overall sentiment for any product.

That’s how ComparaSum was born—a tool that helps users quickly compare products and make informed buying decisions without the hassle of endless scrolling.

⚙️ What It Does

ComparaSum is an AI-powered review summarizer that makes shopping faster and easier by:

✅ Summarizing reviews for a given ASIN (Amazon Standard Identification Number)
✅ Highlighting key pros & cons to help with decision-making
✅ Comparing multiple products side by side
✅ Providing overall sentiment analysis based on thousands of reviews

Right now, we’re focused on Beauty & Personal Care products 💄🛁, as handling all of Amazon’s data would be too much. Our data goes up to 2023, ensuring relevant insights.

🏗️ How We Built It

Our system is built using a full-stack approach with LLMs, AWS, and MySQL:

🔹 Backend: Python + Flask to fetch and process data
🔹 Frontend: React + TypeScript + Tailwind CSS for a clean UI
🔹 Database: MySQL RDS storing Amazon review data
🔹 LLM Model: Llama via Groq for fast, high-quality text summarization
🔹 Cloud Infrastructure:

EC2 instance to store and process the raw review dataset
MySQL RDS to manage structured product data

How It Works:

1️⃣ We downloaded a massive dataset of Amazon reviews to an EC2 instance
2️⃣ Data was structured & stored in MySQL RDS
3️⃣ A Flask backend fetches relevant reviews based on ASIN
4️⃣ Our LLM model (Llama via Groq) summarizes the data
5️⃣ The React frontend displays the summarized insights

⚒️ Challenges We Ran Into

🚀 AWS Setup – Setting up EC2, RDS, and networking for the first time was a steep learning curve
🔗 Database Connectivity – Making MySQL RDS work with EC2 & Flask required debugging
🤖 Choosing the Right LLM – We tested multiple models before finding Llama via Groq for the best balance of speed & accuracy
⚡ Time Constraints – Since this was our first hackathon, working under pressure to build a full-stack AI app was a challenge!

💪 Accomplishments That We're Proud Of

✅ Built a working AI-powered summarization tool 🎉
✅ Successfully set up AWS infrastructure from scratch ☁️
✅ Learned to work with EC2, RDS, and MySQL Workbench 🛠️
✅ Created a React + Flask web app with real-time data fetching 🔥
✅ Many of us were first-time hackathon participants, and we delivered a functional product! 🚀

📕 What We Learned

🔹 How to deploy & manage AWS services (EC2, RDS, S3, networking)
🔹 Working with large-scale datasets in MySQL
🔹 Optimizing Flask + React for better performance
🔹 Efficient LLM-based text summarization
🔹 Debugging database & API issues under time pressure

👀 What's Next for ComparaSum

🔗 Accept direct Amazon links, not just ASINs
🗂️ Expand the dataset beyond Beauty & Personal Care
🖼️ Include product images from reviews (already stored in our database)
⚡ Improve LLM efficiency for even faster summaries
📱 Make the UI more mobile-friendly