-
-
Introduction
-
Logging a meal or transport activity
-
Community(Global) Level impact - Leaderboard display to show gamification
-
AI Butler Interaction: Screenshot showing the Butler giving context-aware advice and the interface for photo or PDF uploads.
-
Mobile view of the main dashboard with eco score of user
-
Forest impact from actions
-
Eco-Envisioning simulator with sliders for solar/EV + dark mode
-
Insights graphs screen
Inspiration
Climate change is one of the biggest challenges of our time, and while many people want to reduce their carbon footprint, taking meaningful action can feel overwhelming. Most carbon trackers are just spreadsheets - dry, confusing, and hard to maintain consistently. We wanted to create a solution that makes sustainability tangible, rewarding, and easy to integrate into daily life.
The Carbon Footprint Butler transforms abstract environmental impact into visible, measurable results. By turning every actionl; cycling instead of driving, eating plant-based meals, reducing energy consumption; into growth in a personal Digital Forest, users can see the direct effects of their choices. This creates real behavioral change, helping people reduce CO₂ emissions while building long-term eco-conscious habits.
What it does
The Carbon Footprint Tracker is a full-stack Flutter application that:
- Gamifies Sustainability: Your actions impact a living Digital Forest. Save CO₂, and watch your "Fledgling Woods" evolve into an "Earth Guardian Realm."
- AI Butler Service: An intelligent "Eco Butler" (powered by Google Gemini) provides context-aware briefings, actionable suggestions, and motivates streaks.
- Vision-Powered Logging: Snap a photo of a meal or upload a PDF energy bill. Gemini Vision extracts and logs your carbon impact automatically.
- Eco-Envisioning: A simulator to preview the future. Toggle sliders for solar installations or EV adoption to see annual CO₂ and cost savings.
How we built it
- Backend (Serverpod): Type-safe endpoints for ActionLogging, Leaderboards, and AI Processing. Code generation eliminated manual JSON serialization.
- AI (Google Gemini): The Butler has a "System Persona" that adapts to the time of day and user stats.
- Frontend (Flutter): Custom Glassmorphism design system. Fully responsive from mobile to wide-screen desktop.
- Database (PostgreSQL): Stores user profiles, action logs, and forest states, managed seamlessly by Serverpod ORM.
Scientific Methodology & Math
We use the Avoided Emissions approach aligned with the Greenhouse Gas Protocol:
$$ \text{Impact (kg CO₂e)} = \text{Activity Data} \times \text{Emission Factor (EF)} $$
- Activity Data: Quantitative user action (e.g., “5 km biked”).
- Emission Factor (EF): CO₂e emissions avoided relative to baseline behaviors.
The Eco Score, which drives forest growth, is calculated as:
$$ \text{Eco Score} = \text{Total Avoided Emissions (kg)} \times 10 $$
Challenges we ran into
- Timezone Travel: Butler initially greeted users at the wrong local time. Fixed via client-server TimeContext handshake.
- AI Hallucination: Gemini suggested impossible actions. Solution: constrained suggestions to supported app features only.
- Web Compatibility: dart:io caused upload failures. Solution: switched to byte-stream handling (FileType.any).
Accomplishments that we're proud of
- Butler developed a distinct personality, polite and context understanding, enhancing user experience.
- Silent Sync: Live-updated Eco Score and Leaderboards without loading interruptions.
- Fully responsive Glassmorphism UI across mobile and desktop.
What we learned
- Serverpod Superpower: Instant backend model availability cut development time in half.
- Prompt Engineering is Coding: System Instructions were as critical as Dart code.
- UI Polish Matters: Small touches like glass cards and dynamic forest visuals transformed the app into an immersive experience.
What's next for Carbon Footprint Tracker: Your Digital Eco-Butler
- IoT Integration: Butler will interact with smart plugs and devices.
- Community Challenges: Local carbon-saving competitions.
- AR Forest: Project your digital forest into real-world environments using Flutter AR.
Built with
Flutter, Dart, Serverpod, PostgreSQL, Google Gemini, Render, Docker
Link to deployed project
https://rachelcooray.github.io/carbon-footprint-tracker/
Link to demonstration video
Built With
- dart
- docker
- flutter
- gemini
- postgresql
- render
- serverpod

Log in or sign up for Devpost to join the conversation.