Skip to content

seedance_create_task

Create Seedance video tasks, variations, or a free dry-run preview.

Choose model and limits with ark_list_models(kind=‘video’); invalid requests fail before submission. Regular 2.5 accepts 4K locally; verify provider support. Flash only 480p/720p, without Draft. Draft previews are 480p; finals are 1080p. Cost previews are static. Submit via ark_job_submit, then poll task IDs with gen_task_get; never resubmit to poll, because each create starts new billed tasks.

Requirement Value
Registration Requires BYTEPLUS_MODELARK_API_KEY.
Execution Background required
Required permission seedance:create

This site lists the full configured capability surface, not the tools enabled on your particular server. Use model discovery and job discovery to check your connection. Credentials are server configuration, never tool arguments.

These are supported families and built-in model or service IDs for this site revision, not provider account access or a guarantee that every row is enabled on your server. Check model discovery for configured bindings and service availability.

Model Model or service ID Availability Capabilities
Seedance 2.5 dreamina-seedance-2-5-260628 Built-in default. 480p/720p/1080p/4K on the same model; 4-30 s or auto; 30 images / 10 videos / 10 audios; audio-only references; structured edit/extend; MP4/MOV; Draft final at 1080p, not 4K; Draft window 14 days (7 for creation before 2026-10-28 10:00 UTC+8).
Seedance 2.5 Flash dreamina-seedance-2-5-flash-261115 Added alongside the selected default when bindings are empty; explicit lists must add seedance_2_5_flash. Account activation required; discovery does not verify entitlement. 480p/720p; 4-30 s or auto; 10 images / 5 videos / 5 audios; audio-only references; structured edit/extend; MP4/MOV; no Draft. Static cost previews reuse the 2.5 fallback, not token pricing.
Seedance 2.0 dreamina-seedance-2-0-260128 Opt-in default or configured binding (standard). 480p/720p/1080p/4K; 4-15 s or auto; 9 images / 3 videos / 3 audios; audio references require image/video; no Draft or native edit/extend.
Seedance 2.0 Fast Authorized custom ID; no built-in ID Configure a fast family binding. 480p/720p; 4-15 s or auto; 9 images / 3 videos / 3 audios; audio references require image/video; no Draft or native edit/extend.
Seedance 2.0 Mini Authorized custom ID; no built-in ID Configure a mini family binding. 480p/720p; 4-15 s or auto; 9 images / 3 videos / 3 audios; audio references require image/video; no Draft or native edit/extend.

Pass an exact configured ID as model, or omit it for SEEDANCE_DEFAULT_MODEL. Custom IDs require an explicit family binding in SEEDANCE_MODEL_BINDINGS. Limits below are family limits; discovery is authoritative for your server. Regular 2.5 4K is retained server acceptance, not confirmed by current model-specific provider docs; see the contract caveat.

Ask your connected AI client to use seedance_create_task for the operation you want. The client constructs the arguments and handles tool calls. Review its proposed operation before approving anything billable. For a non-billable introduction, start with discovery without generation.

Submit through ark_job_submit with tool_name and the complete target arguments, including its inner input object. Poll ark_job_get using the returned Ark job_id, respecting poll_after_ms. Inspect both terminal status and result.is_error; the original output is in result.structured_content. Do not call this tool in the foreground.

Exact MCP call (for developers)

The following is an MCP tool-call envelope, not a shell command. Example media URLs and identifiers must be replaced with your own accessible resources. Seedream and Seedance examples are non-billable dry runs; other provider operations may incur charges.

{
"name": "ark_job_submit",
"arguments": {
"input": {
"tool_name": "seedance_create_task",
"arguments": {
"input": {
"prompt": "A paper boat floats on calm water.",
"resolution": "4k",
"dry_run": true
}
}
}
}
}

An Ark job ID is not a provider task ID. Video and 3D tasks use gen_task_get, VOD tasks use vod_get_task, and avatars use omnihuman_get_task. Save accepted provider task IDs and resume polling instead of submitting a second generation.

The MCP arguments object wraps these fields as {"input": {...}}.

Field Type Required Description and constraints
audios array of SeedanceAudioInput or null No Reference audio (regular 2.5: 10; Flash: 5; see ark_list_models). Entry forms as for images. 2.5 and Flash accept audio alone; 2.0 also needs an image or video.; Default: null
draft boolean or null No Draft step 1 (regular 2.5 only): create a 480p preview billed as a normal 480p video, then pass its task_id as draft_task_id. Exclusive with draft_task_id.; Default: null
draft_task_id string or null No Draft step 2: succeeded Draft ID, at most 14 days old (7 if created before 2026-10-28 10:00 UTC+8), to render the final video. Omit prompt, media, duration, ratio, generate_audio and omni_reference_task_type (the provider reuses the Draft’s); resolution must be omitted or the Draft final resolution. Requires variations=1.; Default: null
dry_run boolean No Validate and return the request bodies and a cost estimate; nothing is submitted or billed.; Default: false
duration integer or null No Seconds: -1 for auto, or 4 up to the model maximum (2.0: 15, 2.5: 30). Task type ‘edit’ requires -1 or omitted.; Default: null
execution_expires_after integer or null No Seconds before an unfinished task expires (3600-259200).; Default: null
generate_audio boolean or null No Whether to generate an audio track for the video.; Default: null
images array of SeedanceImageInput or null No Reference images (regular 2.5: 30; Flash: 10; see ark_list_models). Each entry is a URL string, “asset://”, or {“url”, “role”} with role first_frame, last_frame or reference_image (default).; Default: null
model string or null No Exact model ID from ark_list_models(kind=‘video’); the default if omitted. With draft_task_id, omit to reuse the Draft’s model.; Default: null
omni_reference_task_type “auto”, “reference”, “edit”, “extend” or null No Seedance 2.5 task type: ‘auto’ (inferred, default), ‘reference’, ‘edit’ or ‘extend’. ‘edit’ and ‘extend’ need a reference video and ratio omitted or ‘adaptive’; ‘edit’ also needs duration omitted or -1. Write the prompt to match the type or the task fails with TaskTypeMismatch.; Default: null
output_format “mp4”, “mov” or null No Container (2.5 and Flash). ‘mov’ is H.264 4:4:4 with PCM audio for post-production; some players cannot play it.; Default: null
priority integer or null No Task priority (0-9).; Default: null
prompt string or null No Video prompt (1-32,000 characters). Optional when media inputs or variation_prompts are given.; Default: null
ratio “16:9”, “4:3”, “1:1”, “3:4”, “9:16”, “21:9”, “adaptive” or null No Aspect ratio; ‘adaptive’ is the provider default. Seedance 2.5 edit, extend and first/last-frame requests require it omitted or ‘adaptive’.; Default: null
resolution “480p”, “720p”, “1080p”, “4k” or null No Resolution: server accepts all four for regular 2.5; verify provider 4K support. Flash only 480p/720p. Draft preview: 480p; final: 1080p (defaults when omitted). Per-model limits: ark_list_models.; Default: null
return_last_frame boolean or null No Whether to return the last frame as a separate image output.; Default: null
safety_identifier string or null No Content-safety tracking identifier (max 64 characters).; Default: null
variation_prompts array of string or null No One prompt per variation, overriding prompt; exactly variations entries.; Default: null
variations integer No Parallel tasks (1-5). Each is a separate billed task; failures are reported per variation.; minimum: 1; maximum: 5; Default: 1
videos array of SeedanceVideoInput or null No Reference videos (regular 2.5: 10; Flash: 5; see ark_list_models). Entry forms as for images.; Default: null
watermark boolean or null No Whether to apply an AIGC watermark to the video.; Default: false
Field Type Required Description and constraints
dry_run boolean No True when nothing was submitted; see preview.; Default: false
failed integer No Number of submissions that failed.; Default: 0
family string Yes Resolved model family.
model string Yes Resolved model ID.
preview SeedanceCreatePreview or null No Request bodies and cost estimate; set only for a dry run.; Default: null
recommended_poll_after_ms integer or null No Suggested delay in milliseconds before the first gen_task_get poll; null for a dry run.; Default: null
submitted integer No Number of tasks the provider accepted.; Default: 0
tasks array of SeedanceTaskRef No One entry per requested variation, in order; empty for a dry run.

Field types named above refer to these shared definitions. Optional/null fields, enum choices and bounds are retained in the complete schemas below. Schema validation is followed by runtime capability and security checks.

MediaSourceKind

No fields. Pass {}.

SeedanceAudioInput
Field Type Required Description and constraints
data string or null No Base64-encoded media data. Required when kind is ‘base64’.; Default: null
kind MediaSourceKind Yes Whether the media is referenced by URL or Base64.
mime_type string or null No MIME type of the media (e.g. image/png, audio/wav).; Default: null
role “reference_audio” No Role of this input. Always ‘reference_audio’.; Default: “reference_audio”
url string or null No HTTPS URL of the media, or ‘asset://<asset_id>’ for an Active private asset library asset (ark_asset_create). Required when kind is ‘url’.; Default: null
SeedanceCreateTaskInput
Field Type Required Description and constraints
audios array of SeedanceAudioInput or null No Reference audio (regular 2.5: 10; Flash: 5; see ark_list_models). Entry forms as for images. 2.5 and Flash accept audio alone; 2.0 also needs an image or video.; Default: null
draft boolean or null No Draft step 1 (regular 2.5 only): create a 480p preview billed as a normal 480p video, then pass its task_id as draft_task_id. Exclusive with draft_task_id.; Default: null
draft_task_id string or null No Draft step 2: succeeded Draft ID, at most 14 days old (7 if created before 2026-10-28 10:00 UTC+8), to render the final video. Omit prompt, media, duration, ratio, generate_audio and omni_reference_task_type (the provider reuses the Draft’s); resolution must be omitted or the Draft final resolution. Requires variations=1.; Default: null
dry_run boolean No Validate and return the request bodies and a cost estimate; nothing is submitted or billed.; Default: false
duration integer or null No Seconds: -1 for auto, or 4 up to the model maximum (2.0: 15, 2.5: 30). Task type ‘edit’ requires -1 or omitted.; Default: null
execution_expires_after integer or null No Seconds before an unfinished task expires (3600-259200).; Default: null
generate_audio boolean or null No Whether to generate an audio track for the video.; Default: null
images array of SeedanceImageInput or null No Reference images (regular 2.5: 30; Flash: 10; see ark_list_models). Each entry is a URL string, “asset://”, or {“url”, “role”} with role first_frame, last_frame or reference_image (default).; Default: null
model string or null No Exact model ID from ark_list_models(kind=‘video’); the default if omitted. With draft_task_id, omit to reuse the Draft’s model.; Default: null
omni_reference_task_type “auto”, “reference”, “edit”, “extend” or null No Seedance 2.5 task type: ‘auto’ (inferred, default), ‘reference’, ‘edit’ or ‘extend’. ‘edit’ and ‘extend’ need a reference video and ratio omitted or ‘adaptive’; ‘edit’ also needs duration omitted or -1. Write the prompt to match the type or the task fails with TaskTypeMismatch.; Default: null
output_format “mp4”, “mov” or null No Container (2.5 and Flash). ‘mov’ is H.264 4:4:4 with PCM audio for post-production; some players cannot play it.; Default: null
priority integer or null No Task priority (0-9).; Default: null
prompt string or null No Video prompt (1-32,000 characters). Optional when media inputs or variation_prompts are given.; Default: null
ratio “16:9”, “4:3”, “1:1”, “3:4”, “9:16”, “21:9”, “adaptive” or null No Aspect ratio; ‘adaptive’ is the provider default. Seedance 2.5 edit, extend and first/last-frame requests require it omitted or ‘adaptive’.; Default: null
resolution “480p”, “720p”, “1080p”, “4k” or null No Resolution: server accepts all four for regular 2.5; verify provider 4K support. Flash only 480p/720p. Draft preview: 480p; final: 1080p (defaults when omitted). Per-model limits: ark_list_models.; Default: null
return_last_frame boolean or null No Whether to return the last frame as a separate image output.; Default: null
safety_identifier string or null No Content-safety tracking identifier (max 64 characters).; Default: null
variation_prompts array of string or null No One prompt per variation, overriding prompt; exactly variations entries.; Default: null
variations integer No Parallel tasks (1-5). Each is a separate billed task; failures are reported per variation.; minimum: 1; maximum: 5; Default: 1
videos array of SeedanceVideoInput or null No Reference videos (regular 2.5: 10; Flash: 5; see ark_list_models). Entry forms as for images.; Default: null
watermark boolean or null No Whether to apply an AIGC watermark to the video.; Default: false
SeedanceImageInput
Field Type Required Description and constraints
data string or null No Base64-encoded media data. Required when kind is ‘base64’.; Default: null
kind MediaSourceKind Yes Whether the media is referenced by URL or Base64.
mime_type string or null No MIME type of the media (e.g. image/png, audio/wav).; Default: null
role “first_frame”, “last_frame”, “reference_image” or null No Role of this image: first_frame, last_frame, or reference_image. If omitted, provider default applies.; Default: null
url string or null No HTTPS URL of the media, or ‘asset://<asset_id>’ for an Active private asset library asset (ark_asset_create). Required when kind is ‘url’.; Default: null
SeedanceVideoInput
Field Type Required Description and constraints
kind “url” No Media source kind. Always ‘url’ for video references.; Default: “url”
role “reference_video” No Role of this input. Always ‘reference_video’.; Default: “reference_video”
url string Yes HTTPS URL of the reference video, or ‘asset://<asset_id>’ for an Active private Video asset.
Input schema, including the MCP wrapper
{
"$defs": {
"MediaSourceKind": {
"description": "Whether a media source is a URL or inline Base64 data.",
"enum": [
"url",
"base64"
],
"type": "string"
},
"SeedanceAudioInput": {
"description": "Audio reference input for Seedance.",
"properties": {
"data": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Base64-encoded media data. Required when kind is 'base64'."
},
"kind": {
"$ref": "#/$defs/MediaSourceKind",
"description": "Whether the media is referenced by URL or Base64."
},
"mime_type": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "MIME type of the media (e.g. image/png, audio/wav)."
},
"role": {
"const": "reference_audio",
"default": "reference_audio",
"description": "Role of this input. Always 'reference_audio'.",
"type": "string"
},
"url": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "HTTPS URL of the media, or 'asset://<asset_id>' for an Active private asset library asset (ark_asset_create). Required when kind is 'url'."
}
},
"required": [
"kind"
],
"type": "object"
},
"SeedanceCreateTaskInput": {
"description": "Input model for ``seedance_create_task``.",
"properties": {
"audios": {
"anyOf": [
{
"items": {
"$ref": "#/$defs/SeedanceAudioInput"
},
"maxItems": 10,
"type": "array"
},
{
"type": "null"
}
],
"default": null,
"description": "Reference audio (regular 2.5: 10; Flash: 5; see ark_list_models). Entry forms as for images. 2.5 and Flash accept audio alone; 2.0 also needs an image or video."
},
"draft": {
"anyOf": [
{
"type": "boolean"
},
{
"type": "null"
}
],
"default": null,
"description": "Draft step 1 (regular 2.5 only): create a 480p preview billed as a normal 480p video, then pass its task_id as draft_task_id. Exclusive with draft_task_id."
},
"draft_task_id": {
"anyOf": [
{
"minLength": 1,
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Draft step 2: succeeded Draft ID, at most 14 days old (7 if created before 2026-10-28 10:00 UTC+8), to render the final video. Omit prompt, media, duration, ratio, generate_audio and omni_reference_task_type (the provider reuses the Draft's); resolution must be omitted or the Draft final resolution. Requires variations=1."
},
"dry_run": {
"default": false,
"description": "Validate and return the request bodies and a cost estimate; nothing is submitted or billed.",
"type": "boolean"
},
"duration": {
"anyOf": [
{
"maximum": 30,
"minimum": -1,
"type": "integer"
},
{
"type": "null"
}
],
"default": null,
"description": "Seconds: -1 for auto, or 4 up to the model maximum (2.0: 15, 2.5: 30). Task type 'edit' requires -1 or omitted."
},
"execution_expires_after": {
"anyOf": [
{
"maximum": 259200,
"minimum": 3600,
"type": "integer"
},
{
"type": "null"
}
],
"default": null,
"description": "Seconds before an unfinished task expires (3600-259200)."
},
"generate_audio": {
"anyOf": [
{
"type": "boolean"
},
{
"type": "null"
}
],
"default": null,
"description": "Whether to generate an audio track for the video."
},
"images": {
"anyOf": [
{
"items": {
"$ref": "#/$defs/SeedanceImageInput"
},
"maxItems": 30,
"type": "array"
},
{
"type": "null"
}
],
"default": null,
"description": "Reference images (regular 2.5: 30; Flash: 10; see ark_list_models). Each entry is a URL string, \"asset://<id>\", or {\"url\", \"role\"} with role first_frame, last_frame or reference_image (default)."
},
"model": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Exact model ID from ark_list_models(kind='video'); the default if omitted. With draft_task_id, omit to reuse the Draft's model."
},
"omni_reference_task_type": {
"anyOf": [
{
"enum": [
"auto",
"reference",
"edit",
"extend"
],
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Seedance 2.5 task type: 'auto' (inferred, default), 'reference', 'edit' or 'extend'. 'edit' and 'extend' need a reference video and ratio omitted or 'adaptive'; 'edit' also needs duration omitted or -1. Write the prompt to match the type or the task fails with TaskTypeMismatch."
},
"output_format": {
"anyOf": [
{
"enum": [
"mp4",
"mov"
],
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Container (2.5 and Flash). 'mov' is H.264 4:4:4 with PCM audio for post-production; some players cannot play it."
},
"priority": {
"anyOf": [
{
"maximum": 9,
"minimum": 0,
"type": "integer"
},
{
"type": "null"
}
],
"default": null,
"description": "Task priority (0-9)."
},
"prompt": {
"anyOf": [
{
"maxLength": 32000,
"minLength": 1,
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Video prompt (1-32,000 characters). Optional when media inputs or variation_prompts are given."
},
"ratio": {
"anyOf": [
{
"enum": [
"16:9",
"4:3",
"1:1",
"3:4",
"9:16",
"21:9",
"adaptive"
],
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Aspect ratio; 'adaptive' is the provider default. Seedance 2.5 edit, extend and first/last-frame requests require it omitted or 'adaptive'."
},
"resolution": {
"anyOf": [
{
"enum": [
"480p",
"720p",
"1080p",
"4k"
],
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Resolution: server accepts all four for regular 2.5; verify provider 4K support. Flash only 480p/720p. Draft preview: 480p; final: 1080p (defaults when omitted). Per-model limits: ark_list_models."
},
"return_last_frame": {
"anyOf": [
{
"type": "boolean"
},
{
"type": "null"
}
],
"default": null,
"description": "Whether to return the last frame as a separate image output."
},
"safety_identifier": {
"anyOf": [
{
"maxLength": 64,
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Content-safety tracking identifier (max 64 characters)."
},
"variation_prompts": {
"anyOf": [
{
"items": {
"maxLength": 32000,
"minLength": 1,
"type": "string"
},
"type": "array"
},
{
"type": "null"
}
],
"default": null,
"description": "One prompt per variation, overriding prompt; exactly `variations` entries."
},
"variations": {
"default": 1,
"description": "Parallel tasks (1-5). Each is a separate billed task; failures are reported per variation.",
"maximum": 5,
"minimum": 1,
"type": "integer"
},
"videos": {
"anyOf": [
{
"items": {
"$ref": "#/$defs/SeedanceVideoInput"
},
"maxItems": 10,
"type": "array"
},
{
"type": "null"
}
],
"default": null,
"description": "Reference videos (regular 2.5: 10; Flash: 5; see ark_list_models). Entry forms as for images."
},
"watermark": {
"anyOf": [
{
"type": "boolean"
},
{
"type": "null"
}
],
"default": false,
"description": "Whether to apply an AIGC watermark to the video."
}
},
"type": "object"
},
"SeedanceImageInput": {
"description": "Image input with an optional role for Seedance.",
"properties": {
"data": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Base64-encoded media data. Required when kind is 'base64'."
},
"kind": {
"$ref": "#/$defs/MediaSourceKind",
"description": "Whether the media is referenced by URL or Base64."
},
"mime_type": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "MIME type of the media (e.g. image/png, audio/wav)."
},
"role": {
"anyOf": [
{
"enum": [
"first_frame",
"last_frame",
"reference_image"
],
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Role of this image: first_frame, last_frame, or reference_image. If omitted, provider default applies."
},
"url": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "HTTPS URL of the media, or 'asset://<asset_id>' for an Active private asset library asset (ark_asset_create). Required when kind is 'url'."
}
},
"required": [
"kind"
],
"type": "object"
},
"SeedanceVideoInput": {
"description": "Video reference input for Seedance.",
"properties": {
"kind": {
"const": "url",
"default": "url",
"description": "Media source kind. Always 'url' for video references.",
"type": "string"
},
"role": {
"const": "reference_video",
"default": "reference_video",
"description": "Role of this input. Always 'reference_video'.",
"type": "string"
},
"url": {
"description": "HTTPS URL of the reference video, or 'asset://<asset_id>' for an Active private Video asset.",
"type": "string"
}
},
"required": [
"url"
],
"type": "object"
}
},
"additionalProperties": false,
"properties": {
"input": {
"$ref": "#/$defs/SeedanceCreateTaskInput"
}
},
"required": [
"input"
],
"type": "object"
}
Output schema, including nested definitions
{
"$defs": {
"JsonValue": {},
"SeedanceCreatePreview": {
"description": "What a dry run would have submitted.",
"properties": {
"estimated_cost_usd": {
"description": "Static estimate in USD for all requests; not a provider quote. Does not account for video resolution or duration, including 4K.",
"title": "Estimated Cost Usd",
"type": "number"
},
"requests": {
"description": "Provider request bodies that would be sent, one per variation. HTTPS media URLs are reduced to scheme and host, inline Base64 data to its size, and asset:// references are kept.",
"items": {
"additionalProperties": {
"$ref": "#/$defs/JsonValue"
},
"type": "object"
},
"title": "Requests",
"type": "array"
}
},
"required": [
"requests",
"estimated_cost_usd"
],
"title": "SeedanceCreatePreview",
"type": "object"
},
"SeedanceTaskRef": {
"description": "One submitted Seedance task (or a failed submission).",
"properties": {
"error": {
"anyOf": [
{
"$ref": "#/$defs/VariationError"
},
{
"type": "null"
}
],
"default": null,
"description": "Why this submission failed; null when it was accepted."
},
"index": {
"description": "0-based variation index (0 for a single request).",
"title": "Index",
"type": "integer"
},
"request_id": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Provider request ID for this submission, when returned.",
"title": "Request Id"
},
"task_id": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Provider task ID; poll it with gen_task_get. For draft=true, pass it as draft_task_id once the Draft has succeeded. Null when submission failed.",
"title": "Task Id"
}
},
"required": [
"index"
],
"title": "SeedanceTaskRef",
"type": "object"
},
"VariationError": {
"description": "Machine-readable failure for one variation.",
"properties": {
"ambiguous_completion": {
"default": false,
"description": "Whether the provider may have partially completed despite the error.",
"title": "Ambiguous Completion",
"type": "boolean"
},
"code": {
"description": "Error code (e.g. PROVIDER_ERROR, TIMEOUT, UNEXPECTED_ERROR).",
"title": "Code",
"type": "string"
},
"message": {
"default": "",
"description": "Human-readable error message.",
"title": "Message",
"type": "string"
},
"phase": {
"anyOf": [
{
"enum": [
"queued",
"generating",
"persisting"
],
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Stage at failure: 'queued' (never started, safe to retry), 'generating' (may have completed), or 'persisting' (generated, storing failed).",
"title": "Phase"
},
"request_id": {
"anyOf": [
{
"type": "string"
},
{
"type": "null"
}
],
"default": null,
"description": "Provider request ID, if available.",
"title": "Request Id"
},
"retryable": {
"default": false,
"description": "Whether the caller may retry this variation.",
"title": "Retryable",
"type": "boolean"
}
},
"required": [
"code"
],
"title": "VariationError",
"type": "object"
}
},
"description": "Output model for ``seedance_create_task``.",
"properties": {
"dry_run": {
"default": false,
"description": "True when nothing was submitted; see preview.",
"title": "Dry Run",
"type": "boolean"
},
"failed": {
"default": 0,
"description": "Number of submissions that failed.",
"title": "Failed",
"type": "integer"
},
"family": {
"description": "Resolved model family.",
"title": "Family",
"type": "string"
},
"model": {
"description": "Resolved model ID.",
"title": "Model",
"type": "string"
},
"preview": {
"anyOf": [
{
"$ref": "#/$defs/SeedanceCreatePreview"
},
{
"type": "null"
}
],
"default": null,
"description": "Request bodies and cost estimate; set only for a dry run."
},
"recommended_poll_after_ms": {
"anyOf": [
{
"type": "integer"
},
{
"type": "null"
}
],
"default": null,
"description": "Suggested delay in milliseconds before the first gen_task_get poll; null for a dry run.",
"title": "Recommended Poll After Ms"
},
"submitted": {
"default": 0,
"description": "Number of tasks the provider accepted.",
"title": "Submitted",
"type": "integer"
},
"tasks": {
"description": "One entry per requested variation, in order; empty for a dry run.",
"items": {
"$ref": "#/$defs/SeedanceTaskRef"
},
"title": "Tasks",
"type": "array"
}
},
"required": [
"model",
"family"
],
"title": "SeedanceCreateTaskOutput",
"type": "object"
}
  • Missing tool: check provider credentials and feature flags, then restart the connection.
  • Invalid arguments: keep the correct wrapper and follow current model limits from discovery. A valid schema alone does not prove the chosen model supports the combination.
  • A job completes with an error: inspect the original tool result, not only the outer job status. Never automatically resubmit potentially billable work.
  • Local output writes from jobs require server-configured OUTPUT_ROOTS; HTTP cannot write to a client’s local disk.

See configuration, background execution, and artifact delivery.