One OpenAI-compatible endpoint for every frontier model, with the carbon, water and energy of each request measured and returned in the response. Ecoia offsets more than 200% of every call, so your product ships AI that gives back instead of taking.
If your code already calls OpenAI, you are one line away from a carbon-tracked AI API.
Set the base URL to the Ecoia endpoint and swap in your Ecoia key. The request and response shapes match the OpenAI API, so your prompts, streaming and tool calls keep working unchanged.
A unified /chat endpoint routes to GPT, Claude, Gemini and more, and a /image endpoint handles generation. Pick a model with a single field instead of juggling four vendor SDKs and four billing dashboards.
No second request, no nightly batch job. The footprint travels with the completion.
Grams of CO2 attributed to the request, using grid carbon intensity (~0.395 kg CO2/kWh for the US grid) and data-center overhead (PUE ~1.56).
Liters of cooling and generation water for the request, based on water-use factors of roughly 3.4 L per kWh.
Energy is estimated per token or per image by model class, so the numbers reflect the exact work your call did, not a flat average.
A green AI API that produces audit-ready data, not vibes.
It is an AI inference API that measures the carbon, water and energy used by every request and returns those figures alongside your completion. Ecoia is OpenAI-compatible, so you get the same models you already use plus per-request emissions data and automatic offsetting of more than 200% of each call.
Change one line. Point your existing OpenAI SDK at the Ecoia base URL and use your Ecoia key. The /chat and /image endpoints accept the same request shapes, so most teams switch in minutes with no code rewrite.
Every response carries emissions as headers, including x-eco-co2-grams and x-eco-water-liters, so you can log, display or aggregate the footprint of each call without a second request. The full methodology is published on our How it Works page.
Tracking and offsetting add roughly 5% on top of token cost, and that surcharge already includes retiring verified offsets for over 200% of the measured footprint. There are no separate sustainability fees.
Yes. Aggregated carbon, water and energy data is exportable as CSV and PDF for Scope 3 and ESG reporting, including branded offset certificates suitable for audits and investor decks.
Ecoia.ai runs Claude, GPT & Gemini for chat, images and an API, and offsets over 200% of the water usage and carbon emissions your AI creates.