Evergreen Sentinel : Know Your Waste Laws Before You Toss It

🌿 Evergreen Sentinel — AI Sustainability Auditor

This is a submission for Weekend Challenge: Earth Day Edition

What I Built

Describe any item — a plastic bottle, leftover food, an old t-shirt — and Evergreen Sentinel tells you exactly what to do with it based on your local 2026 waste laws and current weather conditions.

Every audit returns:

• A category — Fabric, Food, or Materials
• A Comfort/Spoilage assessment based on live local weather
• A Circular Action grounded in your city's actual waste regulations

Demo

🔗 Live App

Code

(https://github.com/judexify/evergreen-sentinel)

How I Built It

Stack: React + Vite, Supabase and backboard (auth + persistent state), Groq/LLaMA 3.3 (reasoning), Tavily (real-time search)

On every audit, Tavily fetches live weather and 2026 waste laws for the user's city. Both are injected as context into a Groq prompt which returns a structured JSON audit. All results are saved per user in Supabase with RLS full isolation between accounts.

The key insight: sustainability advice shouldn't be generic. What you do with plastic in Lagos is different from Berlin. The app fetches real, location-specific context every time.

Prize Categories

• Best Use of Backboard — user memory (location, eco score, audit history) managed through a getMemory/setMemory interface backed by Supabase