Inspiration

I wanted to create something that makes people care about their daily environmental impact. Many climate apps feel complicated or impersonal, so I built EcoTrack to make sustainability simple, gamified, and data-driven. It shows users how much CO₂ they save when they make greener choices.

What it does

EcoTrack tracks your eco-friendly actions - biking, recycling, or eating plant-based meals - and calculates how much CO₂ you save. It connects to a Fetch.ai agent that analyzes this data and provides personalized insights, helping users see how small actions add up to a big difference.

How I built it

• iOS app with React Native (Expo).
• Firebase Realtime Database for live activity storage and stats.
• A Fetch.ai uAgents mailbox-backed agent running locally, reasoning via ASI:One.
• FastAPI bridge that the app calls on the LAN; Mailbox relays messages to the hosted/registered         
             agent.

Challenges we ran into

• Firebase anonymous auth & rules that still allow writes from the app but keep things locked                          
            down.
• Wiring up uAgents + Mailbox + ASI.
• Trying to run Expo, Python (FastAPI), and agent logs across ports and networks.
• Time pressure as a solo builder.

Accomplishments that we're proud of

• A working AI-powered sustainability tracker built in under 48 hours.
• Verified agent-driven CO₂ calculations (no hardcoded values) flowing from app → agent → app.
• Clean, responsive mobile UI with real-time stats and history.

What we learned

• Building & registering agents with Fetch.ai, and using Mailbox for local → hosted connectivity.
• FastAPI ergonomics for mobile backends and robust error handling.
• Managing environment variables and keeping secrets out of the repo.
• Debugging multi-service, multi-port setups efficiently.

What's next for EcoTrack – Your Personal Carbon Companion

• Host the agent on Agentverse for 24/7 uptime (no local bridge required).
• Expand leaderboards and achievements to drive habit streaks.
• Smarter recommendations using ASI:One (context-aware nudges).
• Android + Web support.
Share this project:

Updates

Submission history