LLM 2 Leaf

Inspiration

Our inspiration came from the growing use of Large Language Models (LLMs) and the lack of awareness about their significant carbon footprint. Many users don't realize the environmental impact of their interactions with AI systems. We wanted to create a solution that not only highlights this issue but also empowers users to take actionable steps toward sustainability.

What it does

LLM 2 Leaf is a platform that tracks the CO₂ emissions generated from each interaction with an LLM. It provides users with a transparent view of their environmental impact and offers them the opportunity to contribute towards tree-planting initiatives to offset their carbon footprint. By integrating with Verdn, we enable users to make a tangible difference in combating climate change.

How we built it

We built LLM 2 Leaf using Vue.js for the frontend, ensuring a friendly and intuitive user interface with the help of Tailwind CSS. For access to the latest LLMs, we integrated Hugging Face. The backend was developed using Express.js, which serves as a RESTful API to handle requests for tree-planting data from Verdn and other application data stored in MongoDB, such as chat histories.

Challenges we ran into

One of the main challenges we faced was getting the Verdn API to work seamlessly with our application. Additionally, our team had limited familiarity with web technologies like Vue.js and JavaScript, which had a steep learning curve. Debugging and ensuring smooth integration between the frontend, backend, and third-party APIs also posed significant hurdles.

Accomplishments that we're proud of

We are proud of successfully integrating MongoDB for storing user data and chat histories, as well as connecting with Hugging Face to dynamically load LLM responses. Another major accomplishment was accurately displaying sustainability metrics to users, enabling them to make informed decisions about their environmental impact and guiding them toward meaningful pledges.

What we learned

Throughout this project, we learned new frameworks and technologies, including Vue.js, Tailwind CSS, Express.js, and MongoDB. We also gained valuable experience in working with third-party APIs like Verdn and Hugging Face. Beyond technical skills, we learned the importance of collaboration, problem-solving, and the impact of technology on the environment.

What's next for LLM 2 Leaf

For our team, the next steps involve further developing our skills in web technologies, particularly in Vue.js and other modern frameworks. We plan to expand the platform by adding more features, such as personalized sustainability goals and gamification elements to encourage user engagement. Ultimately, we aim to make LLM 2 Leaf a global tool for raising awareness and driving action toward reducing the carbon footprint of AI technologies.