Project overview
Recypher AI is an intelligent waste segregation and disposal guidance platform designed to simplify and improve everyday waste management using artificial intelligence and interactive assistance. The project addresses the widespread confusion people face when deciding whether an item is recyclable, organic, or hazardous, which often leads to incorrect disposal, contamination, and environmental pollution.
The system consists of an Android application powered by on-device AI that allows users to scan any waste item using their phone camera. A TensorFlow Lite model instantly classifies the item and provides clear, easy-to-understand disposal instructions. If the item requires responsible disposal, the app displays nearby waste and recycling centers sorted by distance and guides the user to the appropriate location through integrated map navigation.
To further enhance accessibility and real-time support, Recypher AI includes an AI-powered chatbot that acts as a virtual waste management assistant. Users can ask questions such as “Where should I throw this item?” or “Is this recyclable?” and receive instant, simplified responses. This chatbot bridges the knowledge gap for users who may not always rely on scanning and helps educate them on best waste practices.
To ensure real-world accountability, Recypher AI integrates QR-based disposal verification. When a user reaches a disposal center and scans the QR code, their action is confirmed, and they are rewarded with Green Points. These points contribute to a gamified reward system that encourages consistent sustainable behavior and builds long-term environmental responsibility.
The platform also features a Waste Knowledge Guide that educates users about different waste categories such as plastic, glass, organic, e-waste, and hazardous materials. This section provides useful tips, examples, and best practices for correct segregation. A history dashboard and visual analytics help users track their impact, including carbon savings and waste diversion metrics.
Built with Kotlin, Jetpack Compose, TensorFlow Lite, Node.js, MongoDB, Google Maps integration, and an AI chatbot service, Recypher AI combines mobile technology, machine learning, and smart backend systems to create a reliable and scalable solution. It empowers individuals and communities to make informed decisions and actively contribute toward cleaner cities and a more sustainable future.
Recypher AI transforms waste disposal from a confusing task into a guided, educational, interactive, and rewarding experience — promoting responsible living one scan and one conversation at a time.
Inspiration
We were inspired by the everyday confusion people face while disposing of waste — especially with mixed packaging, used batteries, soiled containers, and items that don’t clearly belong to one bin. Despite good intentions, most people guess, leading to contaminated recycling and increased landfill waste. We realised that the problem wasn’t lack of awareness, but lack of instant, accessible guidance at the moment of disposal. This motivated us to create an AI-powered solution that acts as a real-time waste disposal guide, making correct segregation simple, reliable, and rewarding.
What it does
Recypher AI allows users to scan any waste item using their phone camera and instantly classifies it as recyclable, organic, hazardous, or e-waste. The app provides clear disposal instructions and shows nearby waste and recycling centers on a map. Users can navigate to a center, verify their disposal through QR scanning, and earn Green Points as rewards. It also includes an AI chatbot for quick waste-related queries and a Waste Knowledge Guide to educate users on different waste categories and best practices.
How we built it
We built the Android app using Kotlin and Jetpack Compose with CameraX for image capture and Google Maps for location guidance. Waste classification is performed using a TensorFlow Lite model running directly on the device for fast, offline inference. The backend is developed using Node.js and Express, with MongoDB used to store disposal rules, center data, user history, and points. QR-based verification ensures real-world confirmation of disposal actions, and the chatbot layer provides instant assistance for user queries.
Individual contributions
N.A.
Challenges
Handling classification of visually similar waste items
• Optimising TFLite model size for mobile performance
• Accurately mapping disposal rules across categories
• Implementing reliable QR verification for real-world disposal
• Designing an intuitive UI for elderly and non-technical users
• Integrating real-time map services with minimal latency
Accomplishments
Achieved real-time waste classification on-device
• Built a complete scan-to-disposal workflow
• Implemented QR-based disposal validation system
• Successfully integrated AI chatbot for waste guidance
• Created a gamified reward system for eco actions
• Designed a scalable and modular architecture
Learnings
We learned how to effectively combine AI, mobile development, and backend systems into a unified solution serving a real-world problem. The project improved our understanding of edge-based machine learning, data-driven UI design, API architecture, and user-centric sustainability design. We also gained insight into the complexities of waste management systems and the importance of designing technology that adapts to human behaviour.
Next steps
Expanding coverage to more cities with accurate local waste rules
• Partnering with municipal bodies and NGOs for verified centers
• Enhancing model accuracy with larger, diverse datasets
• Adding multilingual and voice-driven support
• Introducing real-world reward partnerships
• Implementing analytics for city-level waste patterns
• Developing a web dashboard for authorities
Links
