SilkDock
APIVideo Generation

Hailuo 2.3

Configurable fields and supported values for hailuo-2.3.

Model ID

hailuo-2.3

Capabilities

CapabilitySupported value
Text-to-videoYes
Image-to-videoYes
Multi-image / reference videoNo
Start/end framesNo
Motion transferNo
Audio inputNo
Generated audioNo
Resolution720p, 1080p
Aspect ratioText uses model default; image mode follows reference image
Duration (seconds)6 or 10

Configurable fields

FieldRequiredNotes
modelYesMust be hailuo-2.3.
promptYesVideo description or transformation instruction.
secondsNoMust use the supported range or fixed values above.
input_referenceConditionalSingle reference image URL; provide for image-to-video and omit for text-to-video.
sizeNoLegacy pixel dimensions. Prefer extra_body.resolution; aspect ratio is model-controlled.
extra_body.resolutionNoSelect one supported resolution above.
extra_body.prompt_optimizerNoPrompt optimization switch. Defaults to true.
extra_body.fast_pretreatmentNoFast preprocessing switch. Defaults to false.
extra_body.aigc_watermarkNoAdd an AIGC watermark. Defaults to false.

Text-to-video uses the model's default aspect ratio. Image-to-video follows input_reference; this model does not accept extra_body.aspect_ratio.

Do not send unsupported audio, frame, reference, or private provider fields. Account availability is determined by the model list API.

Example

{
  "model": "hailuo-2.3",
  "prompt": "A cinematic tracking shot through a coastal city",
  "seconds": 6,
  "extra_body": {
    "resolution": "1080p"
  }
}

Last updated on

On this page