The AI footprint space has mostly been research papers, academic calculators, and internal ESG dashboards — nothing that just tells a regular person like me (or a team) what their daily AI habit actually costs. ByteThirst was created to close that gap. A few things set it apart: - Coverage across 14 platforms (with more to come), including code generation. - Private by architecture, not by promise. Calculations run locally in your browser with a Chrome Extension or in your IDE with a CLI. The free tier requires no account. Prompts, responses, topics, timestamps, and URLs are never stored. - The browser extension supports the major GPT platforms (ChatGPT, Claude, Gemini, Copilot, Perplexity) but also Lovable, Bolt, and Google AI Studio. That last group matters because a single vibe-coding session generates an estimated ~130× the footprint of a text chat — and no one's been measuring it. You can see all 14 platforms supported on our website. - The CLI supports Claude Code, Aider, Cursor, Codex — where the heaviest developer AI usage actually happens, and where every existing estimator is blind. ByteThirst has first-class CLI support. - We have a transparent methodology with cited sources (Epoch AI, OpenAI, EPA eGRID) on our website. As I started building this to be able to see my own estimates, I realized the value this could have for teams, as well. The Teams tier is built for awareness, not surveillance. The default settings share aggregate QueryWeight measurements with admins, not individual usage. Users can opt-in to share their individual account usage with admins and collaborate towards shared usage goals. Enterprise adds Scope 3 reporting for companies that need AI on the sustainability ledger. Happy to answer anything more specific. I'm eager for feedback and especially curious what the community thinks we're missing.

ByteThirst

Know the water, energy & CO₂ cost of every AI prompt

Know the water, energy & CO₂ cost of every AI prompt