Generate short video clips from text prompts using OpenDunes video models.
OpenDunes generates video through a dedicated asynchronous endpoint. You submit a prompt, get a job back immediately, and poll until the clip is ready. Video rendering takes longer than a single request, so it runs as a background job rather than a blocking call.
| Field | Type | Required | Description |
|---|---|---|---|
model | string | Yes | A video-generation model slug — filter the catalog with ?modality=video |
prompt | string | Yes | Text description of the clip to generate |
seconds | integer | No | Requested clip length (1–60, model-dependent maximum) |
The endpoint responds 202 Accepted with a job id and a poll URL. The cost of the requested duration is held from your balance at submission time:
Poll the job until status is terminal. On completed, the response carries a signed URL to the rendered clip (stored on OpenDunes object storage, time-limited). Re-polling the job returns a freshly signed link; the stored clip itself is retained for a limited time (about two weeks), so download anything you want to keep:
status moves through queued → processing → completed or failed.
Video is billed in DA per second of generated video. The cost of the requested seconds is held on submission and settled on completion. A failed or timed-out job releases the full hold and charges nothing.