Image generation API
Create images from a text description, as PNG, JPEG or WebP.
You describe a subject, the service returns an image. Billing is per call in credits, with no base fee and no minimum volume. Size, quality level and output format are yours to choose per call; an idempotency key makes sure a repeated call is not charged twice. Images are produced in the EU, and nothing is used for training.
View the technical referenceHow this service is billed and processed
Billing is per call in credits, not per month and not per seat. There is no base fee and no minimum commitment: a month without a call costs nothing. What a call costs is fixed before it is triggered, as a fixed figure for services with a fixed price, and for all others as a reserved estimate from which the difference is returned once the actual cost is measured.
Every chargeable call accepts an idempotency key. If the connection drops after the service has done its work but before the response arrives, a retry with the same key returns the same response, and does not charge a second time. This is not a goodwill gesture but part of the interface, and it is described in the OpenAPI file.
Processing takes place within the European Union. One data processing agreement covers this service together with every other, you conclude it once rather than separately per provider. Your inputs and the generated results are not used to train models, neither by us nor by the providers behind them.
The service is reachable two ways, and both lead to the same place: through the API with the key you already have, or through MCP from within an AI assistant, without a line of code. Which one is right depends on whether an application is calling or a person.