AquaTrace

AquaTrace estimates the water, energy, and CO2 footprint of your ChatGPT, Claude, and Gemini conversations — right in your browser, calculated locally, never sent anywhere. HOW IT WORKS - Uses a real BPE tokenizer (the same one OpenAI's models use) to count tokens in your prompts and responses - Auto-detects which specific model you're using (GPT-4o vs GPT-5, Claude Sonnet vs Opus, Gemini Flash vs Pro) - Calculates energy use per model, then converts to water using the actual hosting company's reported water efficiency (AWS, Microsoft Azure, or Google) - Tracks each site separately — nothing gets mixed together or double-counted 100% LOCAL & PRIVATE All calculation happens on your device. No conversation text, token counts, or usage stats are ever sent anywhere — not to us, not to anyone. AquaTrace makes zero network requests. HONEST ABOUT LIMITS No AI company publishes official energy-per-token figures, so every number shown is a clearly-labeled estimate built from published research — never presented as an exact measurement. Full methodology is one tap away in the extension. Works on: ChatGPT (chatgpt.com), Claude (claude.ai), Gemini (gemini.google.com)