EcoSwap

Inspiration

Plastic pollution is one of those problems that feels overwhelming, but it is also driven by thousands of small daily choices. We wanted to build something that helps people make better decisions in the moment, right when they are about to buy a product. EcoSwap was inspired by the idea that sustainability should be easy, practical, and actionable, not something that requires hours of research.

What it does

EcoSwap helps users replace common plastic-heavy products with more sustainable alternatives. A user can search for an item like a plastic bottle, upload an image, or scan a barcode, and EcoSwap immediately suggests eco-friendlier products with direct purchase links. The platform also highlights sustainability-related details, estimated savings, and personal impact over time, allowing users to visualize their contribution to the environment and see that small swaps can add up.

How we built it

We built EcoSwap with a React frontend and a Node/Express backend. The frontend handles the user experience, including product search, image upload, barcode scanning, an impact dashboard, and a chatbot-style assistant. The backend handles product search logic and live product retrieval, using Gemini with web-grounded search to find real alternatives and validate links before returning them to the client. We also added a faster demo path for common searches like plastic bottles to keep the experience responsive during presentations.

Challenges we ran into

One of the biggest challenges was reliability. It is easy to generate suggestions, but much harder to return real products with valid links that users can actually click. We also had to deal with latency from live search, prompt tuning for broad product categories, and making sure sensitive API keys stayed secure by moving model calls to the backend instead of the browser. Balancing speed, relevance, and trustworthiness was a major technical challenge.

Accomplishments that we're proud of

We are proud that EcoSwap goes beyond being a concept and actually connects users to real products they can buy. We built an experience that feels consumer-friendly, with search, barcode scanning, image input, live recommendations, and a personal sustainability dashboard all in one place. We are also proud of the backend improvements that made the app more secure and much more practical for a live demo.

What we learned

We learned that building AI-powered consumer tools is not just about model output, it is about validation, usability, and trust. We also learned how important backend architecture is when working with APIs, especially for protecting keys and improving reliability. On the product side, we learned that users want recommendations that are not only sustainable, but also affordable, convenient, and immediately actionable.

What's next for EcoSwap

Next, we want to improve recommendation quality across more product categories, expand the set of supported retailers, and make search results faster and more consistent. We also want to personalize recommendations based on a user's budget and preferences, add better impact tracking, and integrate directly into shopping workflows so EcoSwap can assist users while they browse online. Long term, we see EcoSwap becoming a smart sustainability layer for everyday shopping.

Tools we used in the process: React Google AI Studio Claude Code Cursor AI VS Code ChatGPT / Gemini for brainstorming