01
The Problem.
AI users have little to no visibility into the energy consumption and carbon footprint of AI tools. This lack of awareness prevents informed decisions and hides the true environmental impact of LLM interactions in everyday use.
02
The Solution.
A multi-platform parser detects each AI provider's text markers, extracts turns and tokens, and converts them into CO₂ equivalents using real energy coefficients — without ever storing raw content. Visualizing comparisons and real-world scenarios makes users aware of their digital footprint in concrete terms.
3
AI Providers
ChatGPT, Claude, Gemini
Tokens → kWh → CO₂
Full conversion chain
Real-world data
Energy coefficients
Infomaniak D4 coefficients
The Stack.
01
Python / Flask
Backend & REST API
02
SQLite
Database management
03
Chart.js
Data visualisation
04
HTML / CSS / JS
Frontend