Inspiration

Sustainability often feels out of reach. Recycling is complicated, and most of our waste still ends up in landfills. In fact, 91% of plastic never gets recycled. We wanted to create something simple, practical, and fun that empowers people, especially women, to turn everyday waste into creative opportunities. T

What it does

UpcycleHer is an AI-powered web app that identifies five common waste materials plastic, glass, cardboard, metal, and batteries in any waste using a computer vision model. Users can upload images of their household waste, and within seconds, the app suggests multiple DIY upcycling projects catered toward the exact item. Each idea comes with step-by-step instructions, difficulty levels, and community ratings, helping people repurpose waste into useful, beautiful creations while joining a community of changemakers.

How we built it

We collected and annotated a dataset of over 2,000 images of household waste using Roboflow, then trained a detection model on YOLOv8. The backend securely calls the Roboflow API, while the frontend was built with Next.js and deployed via Vercel. We designed a clean, efficient UI with Projects and Community tabs to allow users to share their creations and inspire others.

Challenges we ran into

Our computer vision model worked in Colab, but integrating it into our app caused backend/API issues.

Accomplishments that we're proud of

What we learned

We learned how to train and deploy a computer vision model with Roboflow + YOLOv8.

What's next for UpcycleHer

We plan to solidify our computer vision model and deployment and expand our dataset to include textiles, wood, and e-waste, making the tool more comprehensive.