Skip to main content
The API uses the same credits as the app. There is no separate API plan or metering.

​
See the price first

GET /v1/models/{id}/estimate returns the credits a request would charge for the settings you pass (duration, resolution, quality, number of outputs, references), after any discount your organization has, plus whether your balance covers it and whether your plan allows the model. An estimate reserves nothing.

​
When credits move

  • Submit: the estimated amount is reserved.
  • Completed: the reservation becomes the charge; credits_charged on the generation is final.
  • Failed: the reservation is returned. Nothing is charged for a generation that did not finish.
GET /v1/credits shows the balance available to the key right now.

​
Not enough credits

A submit that cannot be covered returns 402 payment_required before anything runs. Top up or change plan in the app under Settings → Billing; pay-as-you-go credits work with the API too.

​
Per-member caps

Workspace admins can cap how many credits each member spends per billing cycle. A key bills as the person who created it, so that person’s cap applies to every request made with the key. GET /v1/credits reports the cap, how much of it is used and when it resets under member_limit when one is in effect. If an integration needs more headroom, an admin can raise or exempt the creator’s cap, or create the key themselves.