Skip to main content
GET
Query Task Status

Authorizations

Authorization
string
header
required

##All APIs require Bearer Token authentication##

Get API Key:

Visit API Key Management Page to get your API Key

Add to request header:

Path Parameters

task_id
string
required

Task ID, ignore {} when querying, append the id from the async task response body at the end of the path

Response

Task status details

created
integer

Task creation timestamp

Example:

1756817821

id
string

Task ID

Example:

"task-unified-1756817821-4x3rx6ny"

model
string

Model used

Example:

"gemini-3-pro-image-preview"

object
enum<string>

Task type

Available options:
image.generation.task,
video.generation.task,
audio.generation.task
Example:

"image.generation.task"

progress
integer

Task progress percentage

Required range: 0 <= x <= 100
Example:

100

status
enum<string>

Task status. Note: while the task is still pending upstream the query API already reports "processing", so clients normally only see processing / completed / failed.

Available options:
pending,
processing,
completed,
failed
Example:

"completed"

type
enum<string>

Task type

Available options:
image,
video,
audio,
text
Example:

"image"

duration
integer

Task duration in seconds. Only present when status is "completed" or "failed".

Example:

30

video_duration
number

Actual generated video length in seconds. Only present for some video models (e.g. Kling) in a final state.

Example:

5

draft_expires_at
string<date-time>

Deadline for conversion to a final video (UTC, RFC 3339). Returned only for a Seedance 2.5 draft (created with draft: true) when its status is completed. After this time, it can no longer be converted using Seedance 2.5 Draft-to-Video

Example:

"2026-09-23T17:19:36Z"

source_task_id
string

ID of the draft task from which this final video was converted. Returned only for Seedance 2.5 Draft-to-Video tasks

Example:

"task-unified-1774857405-abc123"

results
string<uri>[]

List of result URLs. Only present when status is "completed" and the model produces URL results; the field is absent otherwise.

Example:
result_data
any

Model-specific full result object. For image models it is an array of objects such as [{"url": ...}] (same URLs as results); some models return richer objects (e.g. Suno returns song objects). Absent when the model has no structured result.

Example:
usage
object

Usage and billing information. Only present when status is "completed". Always contains credits_used and cost; token fields appear only for token-billed models (e.g. gpt-image-2). Some models add further model-specific fields.

error
object | null

Error information when the task fails (only present when status is "failed"). Note: error.code here is a string-type business error code, different from HTTP status codes. See the Error Codes Reference for the complete list.

task_info
object

Task detailed information

effective_speed
enum<string>

Midjourney V7 variation / pan / outpaint: prompt is optional. Missing, empty, whitespace-only, or controlled-speed/version-flags-only prompts always use Fast and the Fast price, including requests with speed=turbo. With non-empty prompt text, omitted speed or fast uses Fast; turbo uses Turbo at twice the Fast price. Playground offers Fast only. The source must be completed and owned by the current user; promptless operations support image-generation and supported variation, pan, and outpaint results. Historical tasks retain their recorded execution speed and billing.

Available options:
draft,
fast,
turbo