Developer API

LLM carbon footprint API

Estimate and track the CO2e of AI usage from model names and token counts. Built for product teams that need transparent internal reporting without sending prompt text.

No prompt content collected
Chrome, Edge and Firefox
MDM deployment
Token metadata only

Track emissions without collecting text

Send the same metadata your application already receives from LLM providers: model, prompt tokens, completion tokens and optional tenant identifiers.

POST /api/v1/estimate
{
  "model": "gpt-4o",
  "prompt_tokens": 450,
  "completion_tokens": 120
}
POST /api/v1/track
{
  "model": "claude-3-5-sonnet",
  "prompt_tokens": 900,
  "completion_tokens": 280,
  "tenant_id": "customer_123"
}

What teams use it for

Product analytics

Add CO2e next to cost and latency when you monitor AI features in production.

Customer-level views

Use stable tenant identifiers to understand aggregate impact by workspace, customer or project.

Sustainability reporting

Give sustainability teams an explainable view of AI usage and its limits, backed by a public methodology.