ReCycleLink

ReCycleLink — Where AI meets sustainability, through Connecting, collecting, recycling

Comment

Inspiration

ReCycleLink began with a simple realization: although urban households in Bangladesh produce tons of reusable plastic and old clothes every year, most of it ends up in landfills due to the absence of an organized digital recycling system.

We discovered that:

Many people want to recycle but lack access or awareness

NGOs and collectors struggle to coordinate efficiently

Urban areas face increasing environmental pressure

This motivated us to build a digital solution that makes recycling effortless, rewarding, and community-driven—empowering people to participate in a circular, sustainable economy.

What We Learned

Throughout development, we gained crucial technical and real-life insights:

AI & Algorithms

Built waste image classification using pre-trained AI models

Explored optimization algorithms for route planning

Full-Stack Engineering

Developed a scalable system with React, Next.js, Supabase, MongoDB

Human-Centered Design

Studied how gamification, clarity, and rewards improve user engagement

Sustainability Workflows

Analyzed NGO recycling operations and digitized the workflow

These learnings helped us build a solution that is technically robust and socially impactful.

How We Built It

Tech Stack

Frontend: React, Next.js, TailwindCSS

Backend: Supabase (Auth + DB + API), MongoDB

AI Models: Replicate / Hugging Face / TensorFlow.js

Maps & Routing: Google Maps API

Storage: Cloudinary / Firebase

Deployment: Vercel

Core Features

1. Smart Pickup Scheduling

Users select pickup time slots; collectors receive optimized routes.

2. AI-Powered Material Classification

Images are categorized into plastic, metal, glass, paper, fabric, etc.

3. Gamified EcoPoints System

Points encourage consistent recycling behavior:

$$ \text{EcoPoints} = 10 \times (\text{kg}) + 5 \times (\text{verified donations}) $$

4. NGO & Collector Dashboard

NGOs and collectors track pickups, inventory, user activity, and environmental impact.

System Architecture (High-Level)

Next.js → UI + API routes

Supabase → OTP authentication, database, row-level security

MongoDB → Pickup logs, reward histories, activity tracking

Google Maps API → Real-time optimized routing

AI Models → Waste classification + reward recommendation

Challenges We Ran Into

1. No Local Waste Dataset

We created our own labeled dataset to improve accuracy for Bangladesh-specific waste.

2. Route Optimization in Real Conditions

Dhaka/Sylhet traffic and time constraints required heuristic-based refinement.

83. User Engagement

We experimented with:

Gamification

Points

Badges

Leaderboards

Motivational UI triggers

4. Multi-Platform Synchronization

Maintaining consistency across Supabase, MongoDB, and external AI APIs was complex.

5. Understanding NGO Operations

We interviewed NGOs, mapped real workflows, and built digital equivalents.

Accomplishments We’re Proud Of

Fully functional AI-powered waste classifier

End-to-end pickup scheduling system

Gamified recycling model with reward logic

Dashboard tailored for NGOs and collectors

A solution designed specifically for Bangladesh’s urban ecosystem

What’s Next for ReCycleLink

Expand to more cities across Bangladesh

Add carbon footprint analytics for users

Introduce a marketplace for recycled materials

Launch partnerships with schools & corporate CSR programs

Build community recycling challenges

Final Reflection

ReCycleLink is more than just an app—it is a step toward a cleaner, smarter, and more responsible Bangladesh.

Our mission is to:

Reduce recyclable waste entering landfills

Support NGOs with better visibility and digital automation

Create micro-job opportunities for collectors

Empower citizens to develop sustainable habits

ReCycleLink is scalable, practical, and ready for nationwide expansion.

Built With

Share this project:

Updates