GET

/v1/schemas/byteplus/video/contents/generations/tasks

Accept:

kind inputactivity videohash 7d6a36d3…extracted Aug 28, 2026version currentsource upstream spec ↗

field	recorded source
schema	Source linked above. generated Source trace `e3379046f28d4628702fa094c1ce8678f3e103e6579a5dad417f19e5c90f67e3` { "fetchedAt": 1787893860, "extractorVersion": "1", "verifiedAt": null }

This call returns a task id. The video is on GET /contents/generations/tasks/{id}. Poll until status is succeeded, then download content.video_url (valid 24 hours).

Request body

19 properties · 2 required · 6 $defs

property	type	description
model*	string	Seedance model id (e.g. seedance-1-5-pro-251215; the 2.0 family requires the dreamina- prefix) or a preconfigured endpoint id.
content*	CreateContentGenerationContentItem[]	Prompt text plus image / video / audio inputs. The API sorts requests into task types from the roles present — i2v (first frame), flf2v (first + last frame) and r2v (reference media) — and the two families are mutually exclusive.
safety_identifier	string	Opaque per-end-user identifier for abuse attribution, max 64 chars.
callback_url	string	URL that receives a POST with the task payload on each status change.
return_last_frame	boolean	Also return the final frame as a PNG. Default false.
service_tier	string	'default' (online) or 'flex' (offline batch, half price). Note flex tasks are missing from the default task listing unless filter.service_tier=flex is passed.
execution_expires_after	integer	Seconds from created_at after which the task is marked 'expired'.
priority	integer
generate_audio	boolean	Generate a synchronized audio track. Default false.
draft	boolean	Cheap low-fidelity preview render. Default false.
camera_fixed	boolean	Appends a 'fix the camera' instruction to the prompt. Default false.
watermark	boolean	Burn a watermark into the output. Default false.
seed	integer	Randomness seed; integers in [-1, 2^32-1], where -1 means unseeded.
resolution	string	Resolution tier: 480p/720p on Seedance 2.5 and 2.0-fast/-mini, up to 1080p on the 1.x models and 4k on dreamina-seedance-2-0-260128. Matched case-insensitively by the API; there is no 2K tier on any model.
ratio	string	Output aspect ratio: 16:9, 9:16, 4:3, 3:4, 1:1, 21:9, or 'adaptive' (follows the input frame; image-to-video only).
output_format	string
omni_reference_task_type	string
duration	integer	Whole seconds of output (2-12s on 1.0, 4-12s on 1.5-pro, 4-15s on 2.0, 4-30s on 2.5). -1 lets the model choose (2.5 / 2.0 / 1.5-pro).
frames	integer	Frame count, an alternative to duration that allows fractional seconds.

Types

6 $defs

AudioUrlobject

url*

string

ContentGenerationContentItemTypeenum

text | image_url | audio_url | video_url | draft_task

CreateContentGenerationContentItemobject

type*	enum	text \| image_url \| audio_url \| video_url \| draft_task
text	string
image_url	ImageURL
audio_url	AudioUrl
video_url	VideoUrl
role	string
draft_task	DraftTask

DraftTaskobject

id*

string

ImageURLobject

url*

string

VideoUrlobject

url*

string

Validate a payload

before you spend tokens

shell

$ curl -X POST https://modelschemas.com/v1/validate \
    -d '{"provider":"byteplus","endpointId":"contents/generations/tasks","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}