Skip to main content
Like T2V, the output dimension is the explicit size field (full pixel dimensions) — it sets both the aspect ratio and the price tier. At least one of image_urls or video_urls should be supplied — the whole point of this task is to bias generation on references.

Endpoint

Async — returns {id, status: "pending"}; poll GET /v1/videos/generations/{id} until success or failed.

Parameters

string
required
Text prompt describing the desired video. Max 1500 characters.
string[]
Up to 5 reference image URLs (HTTPS, PNG / JPEG).
string[]
Up to 3 reference video URLs (HTTPS, MP4).
enum
default:"\"1920*1080\""
Full pixel dimensions — sets the aspect ratio and the price tier. One of 1920*1080, 1080*1920, 1440*1440, 1632*1248, 1248*1632 (~2M px, 0.15/s),or1280720,7201280,960960,1088832,8321088( 0.9Mpx,0.15/s**), or `1280*720`, `720*1280`, `960*960`, `1088*832`, `832*1088` (~0.9M px, **0.10/s).
integer
default:"5"
Length of the generated clip in seconds. Range [2, 10].
"single" | "multi"
default:"\"single\""
multi lets the model cut between shots; single keeps a continuous take.
boolean
default:"true"
Whether to generate a synchronized audio track.
string
Content to steer the model away from. Max 500 characters.
boolean
default:"false"
Whether to overlay an upstream watermark on the output.
integer
Random-generator initialiser. Range [0, 2147483647].

Example

Poll until terminal:

Reference