Your AI-powered sustainable shopping sidekick that calculates real-time carbon footprints.
💡 Inspiration
Online shopping is incredibly convenient, but its environmental impact is often invisible to consumers. Most shoppers want to make eco-friendly choices, but checking the sustainability, shipping emissions, and supply chain ethics of every single product is exhausting and time-consuming. We wanted to build a seamless tool that surfaces this data instantly right when a buyer is making a decision.
⚙️ What it does
EcoCart AI is a browser extension that analyzes product pages in real-time to score their environmental impact.
• Eco-Scoring System: Grades products from A to F based on material sustainability, manufacturer ethics, and shipping distance.
• Smart Alternatives: Recommends identical or similar products from greener brands or local sellers if the current item has a poor score.
• Carbon Offset at Checkout: Automatically calculates the exact cost to offset the carbon footprint of your delivery and lets you buy offsets with one click.
🛠️ How we built it
We built the prototype during the 48-hour hackathon sprint using a modern, scalable stack:
• Frontend / Extension: JavaScript, HTML, TailwindCSS, and the Chrome Extension V3 API for the overlay UI.
• Backend: Node.js with Express hosted on Vercel to handle API routing.
• AI & Data Extraction: Gemini Pro API to scrape and summarize long product descriptions, material lists, and corporate sustainability reports.
• Database: MongoDB Atlas to cache scored products and reduce external API calls.
🛑 Challenges we ran into
• Scraping Dynamic Pages: E-commerce sites load content dynamically. We initially struggled to capture product details accurately, but solved it by utilizing MutationObservers in our content scripts to wait for the DOM to fully load.
• Context Window Overload: Feeding entire retail webpages into the AI model was slow and expensive. We optimized this by writing a preprocessing function that filters out irrelevant HTML tags, passing only raw product specs and brand metadata to the LLM.
🏆 Accomplishments that we're proud of
• Sub-2-Second Analysis: Optimized our backend and AI prompts to deliver an Eco-Score within 2 seconds of a user opening a product page.
• Beautiful, Non-Intrusive UI: Designed a sleek slide-out panel that blends natively into major e-commerce platforms without disrupting the user experience.
🧠 What we learned
• Chrome Extension Architecture: Gained deep insights into background service workers, content scripts, and secure message passing.
• Prompt Engineering: Learned how to structure strict JSON output schemas from LLMs to ensure our frontend never breaks due to unpredictable text formatting.
🚀 What's next for EcoCart AI
• Support More Retailers: Expand support from Amazon to regional grocery and fashion platforms.
• Gamified Sustainability: Introduce user profiles where shoppers can earn badges, track their lifetime carbon savings, and compete with friends to shop greener.