Skip to content

Bring a still to life

POSThttps://api.vidmoat.com/v1/ai/live-images
Bearer tokenScope ai.video8 or 24 credits per secondTry it in the playground

vidmoat-image-live-1: animate a still. See the model card under Models. Camera moves use depth parallax; ambient, object, face, body and vehicle motion use the generative tier, our own Wan 2.2 model, where it is available. Nothing is added to the picture. Returns a job to poll.

Body#

  • imageobjectrequired
    { "url": "https://..." } or { "uploadId": "..." } (an upload in your account). imageUrl is accepted instead.
  • promptstringrequired
    The motion, up to 500 characters: "slow push-in", "steam rises from the cup", "the sneaker floats".
  • durationSecintegerdefault 5
    Whole seconds, 2 to 10 (the generative tier: 4 or 5). Billed on exactly this number.
  • aspectRatioenumdefault source
    Output is 720p on the short side.
    One of source, 16:9, 9:16, 1:1, 4:5, 5:4, 4:3, 3:4
  • modeenumdefault auto
    Force the class instead of reading it from the prompt. face, body and generative need the generative tier.
    One of auto, camera, ambient, object, face, body, generative
  • moveenum
    Camera move.
    One of push, pull, drift, orbit, pan
  • shotobject
    Optional shot preset, { move?, pace? }: move push_in, pull_out, orbit, pan or truck; pace slow, normal or fast (sets strength). The living image is routed by this choice, not by the prompt words. Other shot moves are refused with the nearest one (use Generate a video for them). Not combined with move or strength.
  • regionstring
    What moves in ambient mode: water, waterfall, waves, clouds, smoke, steam, fire, leaves, wheat, grass, hair, flag, curtain and more.
  • motionenum
    Object motion on the cut-out subject.
    One of float, bob, sway, tilt, drive
  • directionenum
    Direction of the move or flow.
    One of up, down, left, right
  • strengthnumberdefault 1
    0.25 to 2.
  • boxobject
    { x, y, width, height } as fractions of the frame: where the region or subject is.
  • loopbooleandefault false
    Make the end meet the start. One-way moves (push, pull, drift, pan, drive) cannot loop and are refused.
  • seedinteger
    For a repeatable result.
  • projectIdstring
    Project to attach the job to.
  • maxCreditsnumber
    Never spend more than this.
  • quoteOnlyboolean
    true returns the tier, method and price without starting a job.

Response#

Returns 202.

Response fields#

  • idstring
    The job id. Poll it with Get a generation job.
  • statusstring
    PROCESSING.
  • modelstring
    vidmoat-image-live-1.
  • durationSecinteger
    Seconds billed.
  • quoteobject
    { credits, tier, motionClass, method }.
  • actionVerifiedboolean
    Present (false) when the motion asked for could not be confirmed in advance; actionCaveat says why and poseChange names a pose change that was read from the prompt.
  • creditsobject
    { charged, remaining, basis }.
  • pollstring
    The path to poll.
  • quoteOnlyboolean
    With quoteOnly: true the reply is { quoteOnly, model, quote: { credits, perSecond, durationSec, tier, motionClass, method }, available, unavailableReason? } and no job is started.

Errors#

Besides the errors any request can get (authentication, rate limits):

StatusCodeWhen
400invalid_requestNo image or prompt, an unsupported option, a motion only Generate a video can make, or a quote above maxCredits.
404not_foundprojectId is not one of your projects.
402plan_requiredThe account is on Hobby. Living images come with Creator and every plan above it.
402insufficient_creditsThe balance cannot cover the call. Nothing ran. remaining is the balance.
402quota_exceededA living image is already being made on the account.
501not_configuredThe living-image worker is not available on this deployment.
503unavailableThe generative tier has used its GPU time for the month. Camera moves and region cinemagraphs still work; nothing was charged.
422prompt_rejectedThe prompt failed the safety check. Nothing was charged.