Ecosort

Inspiration

Waste sorting systems often fail not because people don’t care, but because the moment of disposal is full of friction: uncertainty, effort, and lack of feedback. We noticed that many existing solutions focus heavily on identifying waste, but very few focus on changing human behavior at the point of action.

Ecosort was inspired by a simple question: what if correct disposal felt immediate, rewarding, and visible? Instead of treating waste sorting as a compliance problem, we wanted to turn it into a behavioral system that encourages better habits while generating meaningful data about where waste actually goes.

What it does

Ecosort is a scan-and-reward system that guides users at the moment they dispose of an item. Users scan an item before throwing it away, receive clear disposal guidance, and earn rewards for correct actions.

Beyond individual guidance, Ecosort creates a data layer that tracks disposal behavior and waste flow over time. This enables insights into contamination patterns, resource leakage, and recycling efficiency — helping reduce waste at both the individual and system level.

Rather than emphasizing AI as the product, Ecosort uses intelligence quietly in the background to reduce friction, reinforce good behavior, and improve outcomes.

How we built it

We built Ecosort as a full-stack web application with a focus on speed, clarity, and real-world usability.

The frontend is built with React and TypeScript, using Tailwind CSS and shadcn/ui for a clean, accessible interface. Animations and transitions are handled with Framer Motion to make interactions feel smooth and responsive.

On the backend, we used Supabase for authentication, data storage, and edge functions. AI services are used to assist with item understanding and categorization, but they are intentionally abstracted away from the user experience. The system prioritizes fast feedback and consistency over technical novelty.

We also implemented analytics and visualization components to surface trends in disposal behavior and waste flow.

Challenges we ran into

One of the biggest challenges was balancing accuracy with usability. Overly complex classification can slow users down, while oversimplification can reduce trust. We iterated heavily on interaction design to keep the experience fast and intuitive.

Another challenge was designing incentives that encourage long-term behavior change rather than short-term gaming. We focused on rewards that reinforce consistency and learning, not just single actions.

Finally, integrating multiple services while keeping latency low required careful architectural decisions, especially for real-time feedback.

Accomplishments that we're proud of

• Designing a system that prioritizes behavior change over technical spectacle
• Building an end-to-end product that is usable, fast, and extensible
• Creating a data-driven approach to understanding waste flow, not just classification

What we learned

We learned that sustainability problems are often human problems first. Technology is most effective when it fades into the background and supports better decisions rather than demanding attention.

We also gained experience designing systems that connect individual actions to larger environmental impact, and learned how important clarity and feedback are in habit-forming products.

We validated Ecosort through continuous manual testing and real-world usage during development. Core user flows — including item scanning, disposal guidance, reward allocation, and administrative verification — were repeatedly tested to ensure reliability and low latency. We also tested common failure and abuse scenarios, such as users scanning items but disposing incorrectly, misclassification cases, and reward gaming. These insights directly informed the design of pending rewards, staff verification workflows, and behavioral safeguards. With Ecosort deployed live, we were able to observe real interactions and iterate quickly on usability, performance, and system logic.

What's next for Ecosort

Next, we want to expand Ecosort beyond individual users by integrating with campuses, offices, and municipalities. This includes richer analytics dashboards, customizable incentive systems, and partnerships with recycling and waste management providers.

Our long-term vision is to make Ecosort a behavioral infrastructure layer for waste systems — helping people, organizations, and cities reduce resource waste through better everyday decisions.

Built With

• css
• framermotion
• gemini2.0flash
• html
• javascript
• openaigpt-4o-mini
• react-router-v6
• react18
• recharts
• shadcn
• supabase
• tailwind-css
• tailwindcss
• tanstackquery
• typescript
• vite
• zod