Credits and usage controls

Paid platform traffic uses the key owner's wallet. The gateway reserves credits for work and settles actual usage according to the applicable model pricing. View available, reserved, and consumed credits in the dashboard. Platform administrators manage credit grants; customer applications cannot grant or transfer credits.

Control a request

Set an explicit output limit supported by the selected model: max_tokens, max_completion_tokens, or max_output_tokens as documented for the endpoint and model. If omitted, the gateway applies the caller's managed output policy and clamps it to model, context, and hard policy limits.

An optional request header can impose an estimate ceiling:

X-The-Code-Max-Credits: <positive decimal ceiling>

The gateway rejects a request before dispatch when its estimate exceeds this ceiling. It is a per-request control, not an account-wide monthly budget. Model limits, credit balance, key status, and account policy are enforced independently.

Monitor usage

Use Dashboard → Usage and key insights to review requests and consumed credits. Dashboard totals report activity; viewing them does not impose a new spending limit. Keep request IDs for troubleshooting without logging sensitive content.

Consult the current model information and dashboard for applicable pricing. For credits or billing questions, email Support@the-code.org.