carbon.ai₂

Framing the Problem

What is the problem you are trying to solve? Who does it affect?
Most people have no idea how much carbon they produce every day — from driving, eating, or using technology. there’s no simple way to visualize your personal impact. this affects everyone, especially younger users who spend hours online without realizing the environmental cost of their digital habits.

Idea Explanation

What is your idea? How does it fix the problem?
carbon.ai₂ turns your daily carbon emissions into something you can see. it uses ai to estimate co₂ from the photos you take, your driving activity, and your digital screen time, and visualizes it all through a simple, black-and-white interface. instead of data tables or reports, users see a single carbon score and a living “aura” that changes as their footprint grows. it makes awareness effortless and visual.

Implementation

How do all the pieces fit together?
the ios app frontend is built in swiftui. when users capture photos, the images are sent to the gemini vision api for recognition and co₂ estimation. results are stored in supabase, which acts as both backend and database. the frontend pulls updated data from supabase to display the total emissions on the home screen. ai summaries and suggestions come from openai’s api. all parts connect through simple api calls to keep the experience smooth and responsive.

Challenges

What did you struggle with? How did you overcome it?
we struggled to combine multiple ai systems into one clean experience without making the app feel complex. calculating realistic co₂ values for different items was also a challenge. we solved it by building a lightweight mapping system and using published carbon datasets for accuracy.

Accomplishments

What did you learn? What did you accomplish?
we learned how to integrate ai vision, databases, and ui into a single functional prototype. we built a full stack app that recognizes objects, logs emissions, and updates visual feedback instantly. our biggest accomplishment was keeping everything minimal and usable while still technically deep.

Next Steps

What are the next steps for your project? How can you improve it?
next, we’ll refine accuracy using apple’s screen time apis to track digital emissions, improve ai suggestions, and add account syncing. longer term, we want carbon.ai₂ to give real-time insights that help users reduce their footprint automatically — simple, fast, and accurate.