The API surface

The ledger needs one thing from you: token metadata. Model names, token counts, regions, timestamps. Never prompts, never completions; content-shaped payloads are rejected at the API boundary. Here is what it connects to today.

Anthropic

available

Managed connection to organization usage reporting. Daily tokens by model, with cached and reasoning splits where reported.

OpenAI

rolling out

Managed connection to organization usage reporting. Daily tokens by model, cached input reported separately.

Google

rolling out

Managed connection for Vertex AI and Gemini API usage. Token metrics by model and region.

Gateways and self-hosted

available

OpenRouter, LiteLLM, Bedrock, and similar: push daily aggregates through the metering API from infrastructure you already run.

Any provider, no code

available

Self-report monthly token totals in the console, straight from your provider's usage dashboard. The path for consumer plans without usage APIs.

Setup lives in the console

The full connection guide, credential scopes, the metering API request format, and per-plan walkthroughs are in the console, available to every account including Free. Create an account, open the console, and the guide walks you through each source.

Connect up to 50 sources per account; contact us beyond that. Connected sources are preferred over self-report: fresher, more granular, and harder to mistype. Whatever the source, retirements accrue only up to your tier's billable-token cap, and certificates state the token basis of every period.

Next: plans sized on billable tokens

See pricing