-
-
Sortify Website Launch Page
-
Object Recognition Description - Website
-
Language Model Overview - Website
-
Data and Scales of Real World - Website
-
Google Maps Bin tracker - Google Maps API - App
-
Global Leaderboard - App
-
Profile View - App
-
Hardware Sync - Coming Soon - App
-
Capture Image Recognition - Gemini API - App
Inspiration
Waste sorting is one of the simplest ways to reduce environmental impact — yet it’s also one of the most confusing. Different cities follow different recycling rules, contamination rates are high, and most people don’t know whether an item should go in trash, recycling, or compost. We were inspired by how much recyclable material is lost due to uncertainty and misinformation. We wanted to create a tool that removes guesswork, educates users in real time, and turns sustainable behavior into an engaging experience. That motivation led to Sortify — an AI-powered waste classification and eco-action platform.
What it does
Sortify helps users correctly dispose of everyday items using AI. Users can: 📸 Scan or upload an image of an item 🤖 Receive instant AI classification (Recycle / Compost / Trash) 📚 Get clear explanations on why an item belongs in a specific category 🗺️ View nearby disposal bins using an interactive map 🏆 Track progress through profiles, rankings, and gamified eco-scores By combining AI accuracy with education and gamification, Sortify encourages long-term sustainable habits, not just one-time actions.
How we built it
Sortify is built as a modern, scalable web application:
- Frontend: React + Vite + TypeScript + Tailwind CSS
- AI: Multimodal image classification using Gemini API
- Backend: FastAPI (Python) with secure authentication
- Database: SQLite / Firebase authentication
- Maps: Google Maps API + OpenStreetMap data
- Design: Figma-based UI focused on clarity and speed
- We designed Sortify with modular APIs, allowing future expansion into hardware integrations, public - - waste systems, and enterprise partnerships.
Challenges we ran into
- AI ambiguity: Some items don’t clearly belong to one category, requiring explanations rather than just labels
- Model confidence: Balancing accuracy with user-friendly explanations
- Geolocation reliability: Matching real-world bin data with map accuracy
- UI clarity: Making AI output understandable without overwhelming users
- Each challenge pushed us to refine both the technical pipeline and the user experience.
Accomplishments that we're proud of
- Successfully built an end-to-end AI waste classification system
- Integrated real-time explanations alongside predictions
- Designed a gamified sustainability experience instead of a static tool
- Built a production-ready architecture suitable for scaling
- Created a solution with real-world environmental impact
What we learned
- Clear explanations matter just as much as AI accuracy
- Sustainability tools need to be engaging, not just informative
- Modular architecture makes rapid iteration possible during hackathons
- Real-world data is messy — designing for uncertainty is essential Most importantly, we learned that small, well-guided actions can create meaningful environmental change.
What's next for Sortify
📱 Mobile app support for faster scanning 🧠 Improved AI confidence scoring and feedback loops 🗑️ Smart bin integrations with fill-level detection 🌐 City-specific disposal rules and regulations 🤝 Partnerships with schools, campuses, and municipalities Our long-term vision is for Sortify to become a global eco-logistics interface — helping people everywhere make better environmental decisions, one item at a time.
Built With
- geminiapi
- google-maps
- html
- python
- react
- sqlite
- tailwindcss
- tensorflow
- typescript
Log in or sign up for Devpost to join the conversation.