Inspiration

The rise of ChatGPT has had a huge carbon impact with each medium length query emitting around 5 times the amount of carbon dioxide equivalent gasses of a normal Google search (this has decreased over the last two years). With over 1 billion queries being made every day, that's A LOT of CO2e!

We decided to tackle this issue and help cut this carbon emission dramatically, and address the issue of many students choosing to use LLMs to answer simple queries, rather than learning to solve a problem themselves. LLMs are basically destroying people's ability to problem solve and think critically (and is especially prevalent in younger generations), and making a step towards fixing that would be incredible.

What it does

Users are blocked from entering ChatGPT queries with their entry being optionally redirected into Ecosia - a carbon neutral search engine.

Users are scored based on the ratio of their GPT->Ecosia conversions, via our handy extension. The less they choose to use ChatGPT, the better they score on CO2 savings!

Users scores are reflected through their own plant avatar, the higher the score, the happier the plant. The amount of carbon they've subsequently saved over time is also displayed to them as a further incentive.

We then also developed a prototype website with a leaderboard, where users can compete with each other.

How we built it

Tech used:

Frontend: JavaScript, React, TypeScript, HTML and CSS

Backend: Python, Flask and SQlite

Challenges we ran into

As browser extensions are stateless it meant we also developed a REST API, written in Python using Flask, that communicates with an SQLite database.

Accomplishments that we're proud of

We managed to produce a feature-complete product involving an integrated frontend, backend and simple user interface.

What we learned

Developing a browser extension was a new challenge for the whole team, due our lack of knowledge in both React and Chrome extensions, so a fair amount of 'learning on the job' throughout the day and night took place.

What's next for Tree-direct

As the backend is currently hosted locally, our next step would be to shift it to a webserver, to better enable user account management and facilitate fully functional leaderboards.

Share this project:

Updates