Gobin - Smart Recycle Scanner

Inspiration

The inspiration for Gobin came from having trouble figuring out if the reusable bottles we were handed out in the beginning could be recycled. But also from witnessing how inaccessible recycling guidelines are for people with disabilities. For individuals with visual impairments, cognitive differences, or motor challenges, small recycling labels, inconsistent symbols, and text-heavy instructions create barriers to proper waste disposal. We recognized that traditional recycling systems exclude millions of people, exacerbating environmental harm. Motivated by the principles of universal design, we aimed to create a tool that democratizes recycling knowledge through multimodal, assistive technology, empowering all users to participate in sustainability efforts, regardless of ability. With recycling rates stagnating worldwide and confusion about what can and cannot be recycled, we realized that technology could bridge this knowledge gap. We were motivated to create a solution that empowers individuals to make informed recycling decisions through accessible technology, ultimately contributing to a more sustainable future.

What it does

Gobin is an AI-powered accessibility-first recycling assistant that helps users determine the recyclability of items through image recognition designed to break down barriers in waste management. Users simply upload a photo of an object, and Gobin identifies the object and its material composition, analyzes its recyclability with a detailed percentage score, provides specific disposal recommendations and tips, and offers greater insights on biodegradability and bio-recycling potential.

How we built it

We built Gobin using the usual modern tech stack: Frontend: Next.js with React for a responsive, component-based UI Styling: Tailwind CSS and ShadCN for clean, efficient styling AI Integration: Google's Gemini 2.0 Flash model for image analysis and recyclability assessment Hosting and DNS: Netlify for hosting and Tech domain for the DNS domain for our project (https://gobin.tech/) Database: MongoDB Atlas for storing scan results and global object history

Challenges we ran into

Building Gobin presented several engineering challenges. One of the larger ones was optimizing the image processing pipeline to provide quick results while maintaining accuracy on mobile devices. Setting up MongoDB Atlas and designing an efficient schema for storing scan results was incredibly tricky because the cluster kept timing out. Thankfully that problem went away when I switched from the guest wifi to the hotspot.

Accomplishments that we're proud of

We are particularly proud of developing a comprehensive recycling app featuring an intuitive UI, a powerful material identification system, and a vast recycling database.

What we learned

This was one of our first times using MongoDB Atlas for building scalable applications and we appreciated the speeds and JSON like approach of the database.

What's next for Gobin - Smart Recycle Scanner

We have ambitious plans to expand Gobin's capabilities. We plan to implement community-driven accessibility through crowdsourced alt-text descriptions for obscure objects, and integrating with smart devices for voice assistant compatibility. We also plan to include location-based recycling guidelines, community features, an expanded material database, offline functionality, a mobile app, gamification, integration with smart bins, and business solutions, to make proper recycling accessible to everyone and contribute to a more sustainable future.

Built With

• atlas
• css
• gemini
• javascript
• jsx
• lucide
• mongodb
• next.js
• react
• shadcn
• tailwind
• tsx
• typescript