ByteDanceSeedance 2.5Text → Video
Latest cinematic video with audio, multi-reference input, and mp4/mov output in 10- or 8-bit. Up to 30s.
Overview
Alongside the unified model APIs, we expose compatibility APIs that take each vendor's original parameters exactly as the vendor defines them — nothing renamed, nothing reshaped on the way through.
That makes them the shortest path onto Picsart if you already work with the vendor directly: the request bodies you've already written keep working as they are. You keep their parameter names and defaults, and you get the vendor's full parameter set rather than the subset that is shared across every model.
The cost is that a request is written for one vendor — switching models later means rewriting it, and results come back in the vendor's own shape. When you'd rather write once and change models freely, use the unified API.
Make a request
Call the workflow with ai.apis.run() in TypeScript, or hit /workflows/{workflow}/execute directly — params is passed through untouched either way.
These endpoints are addressed by workflow name rather than model id: the one in this model's header, since one model runs as one workflow. Authentication is unchanged — your Picsart API key as a bearer token (see Authentication).
import { createClient, ApiRunMode } from '@picsart/ai-sdk';
const ai = createClient({
apiKey: process.env.PICSART_API_KEY,
apiUrl: 'https://api.picsart.com',
});
// Calls the 'seedance' workflow directly — params are sent as-is.
const { result, usage } = await ai.apis.run('seedance', {
content: [
{
type: "text"
}
]
}, {
mode: ApiRunMode.SYNC,
});
console.log(result); // workflow-specific output
console.log(usage?.credits); // amount charged, in $Async (submit & poll)
This model can run longer than the sync limit (~20s). In TypeScript, ai.apis.run(…, { mode: ApiRunMode.ASYNC }) polls for you; over HTTP, submit and poll yourself.
import { createClient, ApiRunMode } from '@picsart/ai-sdk';
const ai = createClient({
apiKey: process.env.PICSART_API_KEY,
apiUrl: 'https://api.picsart.com',
});
// mode: ASYNC submits the job and polls under the hood — you just await.
const { result } = await ai.apis.run('seedance', {
content: [
{
type: "text"
}
]
}, {
mode: ApiRunMode.ASYNC,
});
console.log(result);Parameters
13 parameters, sent inside params. These are the vendor's own names, so they line up one-for-one with the vendor's documentation. Required ones must be supplied; the rest fall back to their defaults.
| Parameter | Type | Required | Default | Details |
|---|---|---|---|---|
modelModel name that will be resolved to the actual endpoint | string | no | seedance_1_0_pro | seedance_1_0_proseedance_1_0_pro_fastseedance_1_5_proseedance_2_0seedance_2_0_fastseedance_2_0_miniseedance_2_5 |
contentInput content items. All models support text and image_url. v2 models also support video_url and audio_url, enabling edit-video and extend-video flows via reference_video role. Limits for v2 (2.0): max 9 images, 3 videos, 3 audios; audio cannot be the only modality. Limits for 2.5: max 30 images, 10 videos, 10 audios; audio-only input is allowed. Roles: first_frame, last_frame (v1+v2), reference_image/reference_video/reference_audio (v2 only). Asset references (asset://<id>) are accepted as URL values for v2. | ContentItem[] | no | — | — |
└ typeType of content item | string | yes | — | textimage_urlvideo_urlaudio_urldraft_task |
└ textText content (only when type is text) | string | no | — | — |
└ image_urlImage URL object (only when type is image_url) | object | no | — | — |
└ urlURL of the image | string | yes | — | — |
└ video_urlVideo URL object (only when type is video_url). Only supported by v2 models. | object | no | — | — |
└ urlURL of the video | string | yes | — | — |
└ audio_urlAudio URL object (only when type is audio_url). Only supported by v2 models. | object | no | — | — |
└ urlURL of the audio | string | yes | — | — |
└ draft_taskDraft task object (only when type is draft_task). seedance_2_5 only: renders the final 1080p video from a previously generated 480p draft, referenced by draft_task.id. The final reuses the draft's prompt, inputs, ratio, duration and audio; anything else sent with it is passed to the vendor as-is, which rejects what a final does not take (other content items, ratio, duration, generate_audio, draft: true, a resolution other than 1080p). resolution defaults to 1080p. | object | no | — | — |
└ idThe id of a succeeded Seedance 2.5 draft, as returned by the draft: true request in its draft_task. Valid for 7 days. Pass the draft_task object back unchanged. | string | yes | — | — |
└ roleRole of the content item (first_frame, last_frame, reference_image, reference_video, reference_audio) | string | no | — | first_framelast_framereference_imagereference_videoreference_audio |
draftDraft mode, step 1 (seedance_2_5 only): generate a cheap 480p preview to check scene structure, shots and motion. resolution defaults to 480p (the vendor rejects any other). Requires options.drive — the draft is saved there (the file backs the reference, with its expiry on the seedance~draftExpiresAt attribute) and a draft without it is rejected. The result carries a draft_task object — pass it back unchanged within 7 days as a single { type: "draft_task", draft_task } content item (step 2) to render the final video with the same model, at 1080p. Billed as a regular 480p generation, and the final at 1080p for the draft's input (if any) plus output seconds. | boolean | no | — | — |
resolutionVideo resolution (480p, 720p, 1080p, 4k). 1080p supported by seedance_1_0_pro, seedance_1_0_pro_fast, seedance_2_0 and seedance_2_5 (NOT seedance_2_0_fast, NOT seedance_2_0_mini). 4k supported by seedance_2_0 only (H.265/HEVC, 10-bit). seedance_2_5 supports 480p, 720p and 1080p. Affects billing. | string | no | — | 480p720p1080p4k |
ratioVideo aspect ratio. "adaptive" lets the model choose based on input. For seedance_2_5, first/last-frame generation and video-to-video editing/extension tasks (a video input) require "adaptive" — the worker forces it for those (output proportions follow the input), so any explicit ratio is ignored there — unless omni_reference_task_type is set to anything but "auto", in which case the ratio is sent as-is. Text-to-video and reference-to-video keep the provided ratio. | string | no | — | 16:94:31:13:49:1621:9adaptive |
output_formatOutput video container format. seedance_2_5 only. "mp4" (default) is the general-purpose H.264/AAC container; "mov" is a professional H.264 4:4:4 / PCM container with higher color fidelity and larger files. "mp4_8bit" is an mp4 guaranteed to be 8-bit H.264: only 1080p comes back from the vendor as 10-bit H.265/HEVC (which most browsers and editors cannot decode), so at 1080p the video is re-encoded — adding processing time to the request — and at every other resolution the value changes nothing, the vendor already delivers 8-bit. | string | no | — | mp4movmp4_8bit |
omni_reference_task_typeTask type hint, seedance_2_5 only. Omni-reference tasks are one of three subtasks: reference-to-video, video editing and video extension, each with its own parameter constraints. "auto" (vendor default) lets the model infer the subtask from the inputs and prompt; a request incompatible with the inferred one fails asynchronously, after the task was created (InvalidParameter.TaskTypeConstraint). An explicit value makes the vendor validate that subtask's constraints at submit, rejecting a non-conforming request immediately; with an explicit value the worker also stops forcing ratio to "adaptive", so the provided ratio goes out as-is and an incompatible one fails at submit. "reference": generate a new video from reference images / videos / audio — no ratio or duration constraint. "edit": edit an original video's visuals or audio — content needs at least one reference_video of 4–30 seconds, ratio must be "adaptive" and duration -1. "extend": extend an original video forward or backward — content needs at least one reference_video and ratio must be "adaptive". | string | no | — | autoreferenceeditextend |
durationDuration in seconds: 3-12 (v1), 4-15 (v2 / 2.0), 4-30 (2.5), or -1 for auto (seedance_1_5_pro and v2 only). Sent to the vendor as-is and billed as-is (charged per second). Input videos are billed too: the usage amount is the total input video duration plus the output duration. When omitted or -1, the vendor picks the length (following the source video for a video input); the output is charged up front at the model's max (an edit's at its source videos' length) and settled to the served length after generation. For seedance_2_5 editing tasks the vendor may require -1. | object | object | no | — | — |
generate_audioInclude synchronized audio. Supported by seedance_1_5_pro and v2 models. v2 defaults to true when omitted; pass false to disable. Affects billing. | boolean | no | — | — |
return_last_frameReturn the last frame of the generated video as a separate image (populates last_frame_url). Supported by all models, including v2 (seedance_2_0 / seedance_2_0_fast / seedance_2_0_mini) — used for last-frame / face-chaining flows. | boolean | no | — | — |
watermarkWhether to add AI generated watermark | boolean | no | false | — |
camerafixedFix the camera during generation. v1 models only, ignored for v2. | boolean | no | — | — |
optionsOptions controlling safety checks and drive integration | object | no | — | — |
└ safety_checksSafety check settings | object | no | — | — |
└ enabledWhether to run content moderation. Defaults to true. | boolean | no | — | — |
└ driveSave result to Picsart Drive | object | no | — | — |
└ nameFile name in Picsart Drive | string | yes | — | — |
└ attributesCustom attributes to attach to the file | object | no | — | — |
└ folderTarget folder in Picsart Drive | object | no | — | — |
└ inputs_transformationInput transformation settings | object | no | — | — |
└ downscale_oversized_imagesWhether to downscale oversized input images. Defaults to false. | boolean | no | — | — |
Response
Over HTTP the output arrives inside a status envelope, at response.result. ai.apis.run() unwraps that envelope for you and resolves to { result, usage } instead. Either way the result itself is the vendor's own shape.
{
"result": {
"video_url": "https://cdn.picsart.com/…/result.mp4",
"last_frame_url": "https://cdn.picsart.com/…/result.jpg",
"output_format": "…",
"draft_task": {},
"mimeType": "video/mp4",
"driveFile": {}
}
}