EcoCred

EcoCred is an innovative web app, built on the Bolt, that empowers Small and Medium-sized Enterprises to measure, manage, and showcase their ESG performance in a gamified environment enhanced by AI.

Comment

Inspiration

Climate action data is often locked inside spreadsheets and siloed departments. Small–mid-size companies want to reduce their footprint, but struggle to measure progress, keep teammates engaged, and justify ROI to leadership.
EcoCred turns dry ESG reporting into a playful, AI-powered experience that motivates action and surfaces insights in real-time.

What it does

  • 📊 Automatically ingests energy, water, waste & scope-3 data and converts it into clear ESG metrics.
  • 🏅 Rewards positive actions with badges, streaks & celebration modals to keep teams motivated.
  • 🤖 Uses Google Gemini to generate contextual AI recommendations, risk alerts and “Explain-like-I’m-5” summaries.
  • 📈 Provides predictive analytics & ROI calculators so finance teams see the dollars saved, not just tonnes of CO₂e.
  • 📱 100 % PWA: works offline, syncs when re-connected, and supports push notifications for nudge-style coaching.
  • 🗣️ Multi-modal interface (voice input, QR codes, image analysis) so field teams can log data in seconds.

How we built it

Front-end: React + TypeScript with Vite, TailwindCSS and custom hooks for offline caching, circuit-breakers and graceful degradation.
Back-end: Supabase (Postgres, Row Level Security, stored procedures, edge functions) plus Deno functions for Stripe billing & push notifications.
AI layer: Google Gemini SDK wrapped in a circuit-breaker to respect rate limits; fallback heuristics ensure uninterrupted UX.
DevOps: Docker, GitHub Actions, Vercel preview deploys and a custom NGINX config for HTTP/2 & service-worker caching.
Analytics: Real-time charts powered by Postgres LISTEN/NOTIFY and web-socket streams.

Challenges we ran into

  1. Gemini quota spikes → solved with an adaptive rate-limiter & exponential backoff.
  2. Complex Supabase migrations → automated SQL change-sets and defensive RLS tests.
  3. Fast-refresh loops in React contexts → refactored providers and introduced an error boundary.
  4. Keeping gamification logic DRY → consolidated badge libraries and added a single useAchievements hook.

Accomplishments we’re proud of

  • End-to-end offline capability with automatic conflict resolution.
  • 30+ database migrations without downtime.
  • Gamified user journey that increased feature-click-through by 47 % in pilot tests.

What we learned

  • The power (and pitfalls) of combining real-time Postgres with client-side caching.
  • How tiny UX tweaks—like celebration modals—can dramatically boost engagement.
  • Writing resilient AI integrations requires treating models like flaky external services.

What’s next

  • Public API so suppliers can push data automatically.
  • Scope-3 estimation models for harder-to-track categories.
  • Open-sourcing the badge library so other climate tools can reuse it.

Built With

Share this project:

Updates