Adaptive bitrate streaming specification.

Packaging format. Value range:

Encryption type.

Playback address.

Media file size, unit: byte.

Digital watermark type. Available values:

List of substream info.

Copyright information.

Digital watermark template id.

Subtitle information list.

Unique identifier for default subtitle.

DRM encryption method.

Adaptive bitrate streaming template ID.

Watermark list. Up to 10 image or text watermarks are supported.

Traceable watermark.

Copyright watermark.

Digital watermark.

External subtitle file. Elements are subtitle IDs. Supports multiple subtitles, up to 16.

Subtitle suppression information list. Up to 2 supported.

Third-party DRM encryption information. It is not currently supported to initiate tasks using third-party DRM info via task flow.

Adaptive bitrate streaming template unique ID.

Template type. Valid values:

Adaptive bitrate streaming template name.

Adaptive bitrate streaming template description.

Adaptive Transcoding Format. Value range:

DRM type. Value range:

Key provider of DRM. Value range:

The SDMC service will gradually phase out subsequently. Please use the VOD DRM encryption service.

Encryption type of DRM. Value range: {"cbcs", "cenc"}

Adaptive Transcoding Input Stream Parameter Information, up to 10 streams.

Whether to prohibit video low bitrate to high bitrate conversion. Value range:

Whether to prohibit video resolution to high resolution conversion. Value range:

Template creation time, in ISO date format.

Last template modification time, in ISO date format.

Segment type.

ts: HLS, internally mapped to ts-segment

fmp4: HLS/DASH, internally mapped to mp4-mp4-segment for HLS and mp4-mp4-byterange for DASH

ts-segment: HLS+TS segment

ts-byterange: HLS+TS byte range

mp4-segment: HLS+MP4 segment

mp4-byterange: HLS+MP4 byte range

ts-packed-audio: HLS+TS+Packed Audio segment

mp4-packed-audio: HLS+MP4+Packed Audio segment

ts-ts-segment: HLS+TS+TS segment

ts-ts-byterange: HLS+TS+TS byte range

mp4-mp4-segment: HLS+MP4+MP4 segment

mp4-mp4-byterange: HLS/DASH+MP4+MP4 byte range

ts-packed-audio-byterange: HLS+TS+Packed Audio byte range

mp4-packed-audio-byterange: HLS+MP4+Packed Audio byte range
Default value: ts-segment. Note: The shard format of adaptive bitrate is based on this field. For DASH, SegmentType can only be mp4-mp4-byterange.

Average segmentation duration

Unit: seconds

Default value: 6

Adaptive bitrate streaming uses SegmentDuration to set the segment duration, not the HlsTime field.

Video parameter information.

Audio parameter information.

Indicates whether to remove the audio stream. Value range:

Indicates whether to remove the video stream. Value range:

Top Speed Codec transcoding parameters.

Audio/video enhancement configurations.

Extended parameter.

Principal ID.

Entity information.

Capability configuration switch. Available values: ON: Enable; OFF: Disable. Default value: ON.

Matting target type: "foreground" / "pattern"

Pattern cutout configuration. This parameter is valid only when Type is pattern.

Matting model selection. This parameter is optional.

Enumeration values:

• auto: automatically choose an appropriate model
• WAND-cutout-1.0-lite: standard version, fastest speed
• WAND-cutout-2.0-lite: enhanced version, faster speed
• WAND-cutout-2.0-flash: enhanced version, quality-speed balance
• WAND-cutout-3.0-lite: enhanced version, faster speed
• WAND-cutout-3.0-flash: enhanced version, quality-speed balance

Operation type. Available values: add, delete, reset. The reset operation will clear the existing face data of the person and add the face data specified by FaceContents.

Human face ID set. This field is required when Type is delete.

Collection of strings after the face image is Base64-encoded. Only jpeg and png image formats are supported.

Figure ID.

Figure name.

Character description.

Face information.

Figure tag.

Application scenarios.

Creation time, in ISO date format.

Last modification time, in ISO date format.

Operation type. Available values: add, delete, reset.

Tag. Length limit: 128 characters.

Keyword.

Keyword tag.

Keyword application scenario.

Creation time, in ISO date format.

Last modification time, in ISO date format.

Keyword. Length limit: 20 characters.

Keyword tag

Dressing model.

Enumeration values:

• WAND-tryon-1.0-lite: Lightweight tier, speed first.
• WAND-tryon-1.0-flash: Balanced tier, balancing effect and latency.
• WAND-tryon-1.0-pro: High-quality tier, effect first.

Input the image list of clothing that need to be replaced. 1-4 images are supported.

Dressing instruction Prompt.

Subject ID.

Subject name.

Main voice ID.

Subject description.

Entity creation time. Format according to the ISO 8601 standard. For details, see ISO date format description.

Input video file type. Valid values:

Media file ID, the globally unique identifier of the file in VOD, assigned by the VOD backend after successful upload. You can get this field in the video upload completion event notification or the VOD console. This parameter is valid when Type is File. Note: 1. An image less than 10M is recommended; 2. Image format values: jpeg, jpg, png.

Accessible file URL. This parameter is valid when Type is Url. Note: 1. An image less than 10M is recommended; 2. Image format values: jpeg, jpg, png.

Storage mode

Enumeration values:

• Temporary: temporary storage. The generated video file will not be stored in VOD. You can obtain a temporary access URL in the event notification, with a valid period of 7 days.
• Permanent: permanent storage. The generated video file will be stored in VOD. You can obtain the FileId in the event notification.

Default value: Temporary

Output filename. It can contain up to 64 characters. By default, the system specifies the filename.

Category ID, used to categorize and manage media. You can create a category through the create category API to obtain the category ID.

Expiry date of the output file. The file will be deleted after this time. The default is permanent with no expiration. The format follows the ISO 8601 standard. For details, see ISO date format description.

Audio duration. Not filled in by default.

Value range: [0, 60]

Output audio format. By default, this is not specified.

Accessible file URL. This parameter is valid when Type is Url. Note: 1. Images less than 7M are recommended. 2. Supported image formats: jpeg, jpg, png, webp.

Media file ID of the audio file, which is the globally unique identifier of the file in VOD, assigned by the VOD backend after successful upload. You can obtain this field in video upload completion event notification or VOD console. This parameter is valid when Type is File.

Audio file URL, which needs to be external network accessible. This parameter is valid when Type is Url.

Input video file type. Valid values:

Media file ID of a video file, which is the globally unique identifier of the file in VOD, assigned by the VOD backend after successful upload. You can obtain this field in the video upload completion event notification or the VOD console. This parameter is valid when Type is File.

Reference video URL, which needs to be external network accessible. This parameter is valid when Type is Url.

Task ID.

Task status. Valid values:

Error code. A non-zero error code is returned when a source error occurs. If 0 is returned, please use the ErrCode of each specific task.

Extended error code. An empty string indicates success, and other values indicate failure.

Enumeration values:

• RequestLimitExceeded: The call exceeds the concurrency limit.
• InvalidParameter.VoilationContent: The input prompt violates the Content Security Policy.
• InvalidParameterValue: Parameter value error.
• FailedOperation: Model tasks are accumulating.
• InternalError: Internal error.
• InvalidParameter: Invalid parameters.
• InvalidParameter.MediaFormat: Invalid media format.
• ContentModerationFailed: Failed to pass content moderation.
• ResourceInsufficient: Insufficient resources.
• ModelGenerateFailed: Failed to generate the model.
• ResourceNotFound: Resource not found.
• OperationCanceled: Operation canceled.
• TaskTimeout: Task timeout.

Error message.

Task progress. Value range: [0-100].

Input information of AIGC audio tasks.

Output information of AIGC audio tasks.

Model name.

Model version.

Scenario type. Values are as follows:

Prompt content for video generation. Supports up to 1,000 characters. This parameter is required when FileInfos is empty.

Prompt content to prevent the model from generating video. Supports up to 1000 characters.

Reference audio content

Reference video info

Whether to optimize Prompt content automatically. When Enabled, the passed-in Prompt will be optimized automatically to enhance generation quality. Valid values:

Output result file output of AIGC image generation.

Additional parameter.

Output audio information

Output video information

Storage mode. Valid values:

Output filename, up to 64 characters. By default, the system specifies the generated filename. Valid when StorageMode is Permanent.

Category ID, used to categorize and manage media. You can create a category and obtain the category ID through the Create Category API. Valid when StorageMode is Permanent.

Expiry date of the output file. The file will be deleted after this time. It is permanent with no expiration by default. The format is based on ISO 8601. For details, see ISO date format description.

File type, for example, mp4 and flv.

Media file playback address.

Media file ID. Valid when StorageMode is Permanent.

Meta-information of the output audio.

Duration

Unit: second

Input file type. Valid values:

File ID, the globally unique identifier of the file in VOD, assigned by the VOD backend after successful upload. You can obtain this field in the video upload completion event notification or the VOD console. This parameter is valid when Type is File. Note: Supported image formats: glb and obj.

Accessible file URL. This parameter is valid when Type is Url. Note: Supported file formats: glb and obj.

Input file type. Valid values:

File ID, the globally unique identifier of the file in VOD, assigned by the VOD backend after successful upload. You can obtain this field in the video upload completion event notification or the VOD console. This parameter is valid when Type is File. Description: Supported image formats: jpg, jpeg, png, bmp, webp.

Accessible file URL. This parameter is valid when Type is Url. Note: Supported image formats: jpg, jpeg, png, bmp, webp.

View image type.

Enumeration values:

• front: front view (required)

• back: back view

• left: left view

• right: right view

• top: top view

• bottom: bottom view

• left_front: left front 45°

• right_front: right front 45°

• Must contain the front view;

• The same ViewType cannot be duplicated.

Storage mode. Valid values:

Input file type. Valid values:

File ID, the globally unique identifier of the file in VOD, assigned by the VOD backend after successful upload. You can obtain this field in the video upload completion event notification or the VOD console. This parameter is valid when Type is File. Description: Supported image formats: jpg, jpeg, png, bmp, webp.

Accessible file URL. This parameter is valid when Type is Url. Note: Supported image formats: jpg, jpeg, png, bmp, webp.

Task ID.

Task status. Valid values:

Error code. A non-zero error code is returned when a source error occurs. If 0 is returned, please use the ErrCode of each specific task.

Expansion error code.

Error message.

Task progress, value ranges from 0 to 100.

Input information of the AIGC Hunyuan 3D task.

Output information of the AIGC Hunyuan 3D task.

Identifier for deduplication. If a request with the same identifier has been sent within the past seven days, an error is returned for the current request. The maximum length is 50 characters. If this is not specified or left empty, deduplication is not performed.

Source context. This is used to pass user request information. The task status change callback returns the value of this field. The maximum length is 1000 characters.

Generate 3D reference image information.

Multi-perspective image information used to generate a 3D model.

The array length must be between 2 and 8 and must contain the front perspective.

Prompt for generating a 3D model.

Generation Type.

Enumeration values:

• Normal: Generate a complete 3D asset (geometry + texture);
• Geometry: Generate only geometry (no texture, faster output speed);
• Texture: Generate only texture (requires filling in MeshInfos)

Reference 3D model used to generate a 3D model.

Whether to enable output PBR material.

Enumeration values:

• Enabled: enable;
• Disabled: disable.

Number of patches. This parameter is valid only when the value of GenerateType is Normal or Geometry.

Whether to retain UV unfolding.

Enumeration values:

• Enabled: reserved;
• Disabled: not retain.

Result format. In addition to the default return of obj and glb, an extra output format.

Enumeration values:

• FBX: FBX format file.

Random Seed. Results can be reproduced with the same Seed input.

Style control word.

Configuration of the output media file of the task.

Output file information for an AIGC Hunyuan 3D task.

Storage mode. Valid values:

Expiry date of the output file. The file will be deleted after this time. By default, it never expires. The format follows the ISO 8601 standard. For details, see ISO date format description.

File type, such as OBJ, GLB, and FBX.

Output file address.

Output file preview address.

Empty if no file preview address is available.

Storage mode

Enumeration values:

• Temporary: temporary storage. The generated video file will not be stored in VOD. A temporary access URL can be obtained in the event notification, with a valid period of 7 days.
• Permanent: retained permanently. The generated video file will be stored in VOD. The FileId can be obtained in the event notification.

Default value: Temporary

Output filename, up to 64 characters. By default, the system assigns the generated filename.

Category ID, used to categorize and manage media. You can create a category and obtain the category ID through the Create Category API.

Expiry date of the output file. The file will be deleted after this time. It never expires by default. The format follows the ISO 8601 standard. For details, see ISO date format description.

Resolution of the generated image. Available values for each model:

• OG: 1K, 2K, 4K. Default: 1K.
• GG 2.5: 1K, 2K, 4K. Default: 1K.
• GG 3.0: 1K, 2K, 4K. Default: 1K.
• GG 3.1: 720P, 1K, 2K, 4K. Default: 1K.
• Kling 2.1: 1k, 2k. Default: 1k.
• Kling 3.0: 1k, 2k. Default: 1k.
• Kling 3.0-Omni: 1k, 2k, 4k. Default: 1k.
• Kling O1: 1k, 2k, 4k. Default: 1k.
• SI 4.0: 1K, 2K, 4K. Default: 1K.
• SI 4.5: 2K, 4K. Default: 2K.
• SI 5.0-lite: 2K, 3K. Default: 2K.
• Vidu q2: 1080p, 2K, 4K. Default: 1080p.
• Hunyuan 3.0: This field is not currently supported. You can set the resolution through the ExtInfo field.
• Qwen 0925: This field is not currently supported. You can set the resolution through the ExtInfo field.

Specify the aspect ratio of the generated image.

• OG: 1:1, 3:2, 2:3, 3:4, 4:3, 16:9, 9:16, 21:9, 9:21;
• GG 2.5: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9;
• GG 3.0: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9;
• GG 3.1: 1:1, 1:4, 1:8, 2:3, 3:2, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 16:9, 21:9;
• Kling 2.1: 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, 21:9;
• Kling 3.0: 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, 21:9;
• Kling 3.0-Omni: 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, 21:9, auto;
• Kling O1: 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, 21:9, auto;
• Vidu q2: 16:9, 9:16, 1:1, 3:4, 4:3, 21:9, 2:3, 3:2;
• SI 4.0: does not support this parameter. You can specify 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, 21:9 via prompt;
• SI 4.5: does not support this parameter. You can specify 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, 21:9 via prompt;
• SI 5.0-lite: does not support this parameter. You can specify 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3, 21:9 via prompt;
• Hunyuan 3.0: unsupported;
• Qwen 2.0: unsupported;
• Qwen 0925: unsupported;

Whether to allow generation of people or human faces. Valid values:

Whether to enable compliance check on the input content. Valid values:

Whether to enable compliance check on output content. Valid values:

Number of images generated. Available values for each model:

• Kling: value range 1-9;
• OG: value range 1-8;

Not effective for other models.

Specify the image file format for model output. If not specified, the model default value is followed. Available values:

• jpeg
• png

Indicates whether to enable the logo watermark. Default value: Disabled. Valid values:

• Enabled: enable;
• Disabled: disable;

AI image generation scenario type. Available values:

• ai_try_on: AI try-on.
• product_image: AI product image generation.
• outpainting: AI outpainting.

The following Type values are abandoned and will no longer be updated:

• change_clothes
• change_clothes_under
• change_clothes_top_wear
• change_clothes_bottom_wear
• change_clothes_full_wear

Required when Type is ai_try_on, indicating the AI dressing configuration parameters.

Abandoned. Please use AiTryOnConfig.When Type is one of the following types, this item is required and represents the AI try-on image generation configuration parameters:

• change_clothes
• change_clothes_under

Required when Type is product_image, indicating the AI-generated product image configuration parameters.

Task ID.

Task status. Valid values:

Error code. A non-zero error code is returned when a source error occurs. If 0 is returned, use the ErrCode of each specific task.

Extended error code. An empty string indicates success, while other values indicate failure.

Enumeration values:

• RequestLimitExceeded: Model invocation exceeded the concurrency limit.
• InvalidParameter.VoilationContent: The input prompt violates the Content Security Policy.
• InvalidParameterValue: Parameter error.
• FailedOperation: Model tasks are accumulating.
• InternalError: Internal error.
• InvalidParameter: Invalid parameter.
• InvalidParameter.MediaFormat: Invalid media format.
• ContentModerationFailed: Failed to pass content moderation.
• ResourceInsufficient: Insufficient resources.
• ModelGenerateFailed: Failed to generate model.
• ResourceNotFound: Resource not found.
• OperationCanceled: Operation canceled.
• TaskTimeout: Task timeout.

Error message.

Task progress, value ranges from 0 to 100.

Input information for the AIGC image generation task.

Output information of the AIGC image generation task.

Identifier for deduplication. If a request with the same identifier has been sent within the past seven days, an error is returned for the current request. The maximum length is 50 characters. If this is not specified or left empty, deduplication is not performed.

Source context. This is used to pass user request information. The task status change callback returns the value of this field. The maximum length is 1000 characters.

Model name.

Model version.

AIGC image generation task input file information.

Prompt content for image generation. When FileInfos is empty, this parameter is required.

Prompt content to prevent the model from generating images.

Whether to optimize Prompt content automatically. When Enabled, the passed-in Prompt will be optimized automatically to enhance generation quality. Valid values:

Generation mode. Valid values:

AIGC image generation output result file output.

Random seed of the model.

Scenario type. Values are as follows:

Pass-through parameter.

Input file type. Valid values:

Media file ID of the image file, which is the globally unique identifier of the file on VOD, assigned by the VOD backend after successful upload. You can obtain this field in the video upload completion event notification or the VOD console. This parameter is valid when Type is File.
Description:

1. Images less than 7M are recommended.
2. Image format values: jpeg, jpg, png, webp.

Accessible file URL. This parameter is valid when Type is Url.
Description:

1. An image less than 7M is recommended.
2. Image format values: jpeg, jpg, png, webp.

Accessible file Base64. This parameter is valid when Type is Base64. Details:

1. The total size of all files must not exceed 7 MB to avoid exceeding the 10 MB upper limit of message size for cloud APIs after conversion to Base64;
2. Image format should be: jpeg, jpg, png, webp;
3. Do not include prefixes such as data:image/jpeg;base64,.

Description of the input image, used to help the model understand the image. Only valid for GEM 2.5 and GEM 3.0.

Valid only when ModelName is OG. Image type.

Enumeration values:

• mask: Image mask.

Information of the output file of an AIGC image generation task.

Storage mode. Valid values:

Output filename, up to 64 characters. By default, the system assigns the generated filename. Valid when StorageMode is Permanent.

Category ID, used to categorize and manage media. You can create a category and obtain the category ID through the Create Category API. Valid when StorageMode is Permanent.

Expiry date of the output file. The file will be deleted after this time. By default, it never expires. The format follows the ISO 8601 standard. For details, see ISO date format description.

File type, such as png or jpg.

Media file playback address.

Media file ID. Valid when StorageMode is Permanent.

Meta-information of the output image.

Quota type

Enumeration values:

• Image: AIGC image task
• Video: AIGC video task
• Text: AIGC text task

Valid only when QuotaLimit is Text. It is used to select the ApiToken that needs a quota limit.

Number of task quotas

Unit:

• When QuotaLimit=Image, the unit is images
• When QuotaLimit=Video, the unit is seconds
• When QuotaLimit=Text, the unit is tokens

Already used amount

Unit:

• When QuotaLimit=Image, the unit is images
• When QuotaLimit=Video, the unit is seconds
• When QuotaLimit=Text, the unit is tokens

Entries per page

scroll_token returned in the previous page response, used for turning to the next page

Detailed data for text generation

Request start time (RFC3339)

Gateway layer request ID.

Dialogue ID returned by the backend model

HTTP status code returned to the client

Model name

Application ID.

Masked api_key: first 8 characters + **** (returned without modification if length ≤ 8)

Whether to return in streaming mode

Input token count

Output token count

Number of tokens for cache hits in prompt cache

Total token count

tokens per second in the generation stage

Time To First Token

Unit: second

End-to-end total time

Unit: seconds

Entry protocol: completions / responses / anthropic

AIGC specification.
Valid values:

Usage data.

Storage mode

Enumeration values:

• Temporary: Temporary storage. The generated video file will not be stored in VOD. You can obtain a temporary access URL from the event notification, with a valid period of 7 days.
• Permanent: Permanent storage. The generated video file will be stored in VOD. You can obtain the FileId from the event notification.

Default value: Temporary

Output media filename, up to 64 characters. By default, the system assigns the generated filename.

Category ID, used to categorize and manage media. You can create a category and obtain the category ID through the Create Category API.

Expiry date of the output file. The file will be deleted after this time. By default, it never expires. The format follows the ISO 8601 standard. For details, see ISO date format description.

Duration of the generated video, in seconds.

Resolution of the generated video.

Aspect ratio of the generated video.

Whether to generate audio. Supported models include GV, OS, Vidu, Jimeng, and Kling.

Enumeration values:

• Enabled: enable
• Disabled: disable

Default value: Disabled

Whether to allow generation of people or human faces. Valid values:

Whether compliance check is enabled for the input. Valid values:

Whether to enable compliance check on output content. Valid values:

Whether to enable video enhancement. Valid values:

1. If the selected resolution exceeds the model's generation resolution, enhancement is enabled by default.
2. For resolutions the model can generate directly, you can also choose direct low-resolution output and use enhancement to obtain the specified resolution.

Whether to enable off-peak. Valid values:

Whether intelligent frame interpolation is enabled for vidu. Valid values:

Indicates whether to enable the logo watermark. Valid values:

Whether to add background music to the generated video.

Enumeration values:

• Enabled: The system will automatically select suitable music from the preset BGM library and add it.
• Disabled: Do not add BGM.

Default value: Disabled

Storage mode. Valid values:

Output media file name, up to 64 characters. By default, the system specifies the generated file name.

Category ID, used to categorize and manage media. You can call the Create Category API to create a category and obtain the category ID.

Expiry date of the output file. Files will be deleted after this time. By default, files never expire. Format according to the ISO 8601 standard. For details, see ISO date format description.

Parameter format: 2025-12-28T00:35:00Z.

Storage mode. Valid values:

Output filename, up to 64 characters. By default, the system specifies the generated filename. Valid when StorageMode is Permanent.

Category ID, used to categorize and manage media. You can create a category and obtain the category ID through the Create Category API. Valid when StorageMode is Permanent.

Expiry date of the output file. The file will be deleted after this time. By default, it never expires. The format follows the ISO 8601 standard. For details, see ISO date format description.

File type, for example, mp4 and flv.

Media file playback address.

Media file ID. Valid when StorageMode is Permanent.

Meta-information of the output video. Valid when StorageMode is Permanent.

Task ID.

Task status. Valid values:

Error code. A non-zero error code is returned when a source error occurs. If 0 is returned, please use the ErrCode of each specific task.

Expansion error code.

Error message.

Task progress. Value range: [0-100].

Input information of the AIGC video conversion task.

Output information of the AIGC video conversion task.

Identifier for deduplication. If a request with the same identifier has been sent within the past seven days, an error is returned for the current request. The maximum length is 50 characters. If this is not specified or left empty, deduplication is not performed.

Source context. This is used to pass user request information. The task status change callback returns the value of this field. The maximum length is 1000 characters.

For description of the style. Limited to 50 characters.

Input file information of the AIGC video conversion task.

AIGC video redrawing task parameter information.

Output configuration for AIGC video conversion.

Input video file type. Valid values:

Media file ID, the globally unique identifier of the file in VOD, assigned by the VOD backend after successful upload. You can obtain this field in the video upload completion event notification or the VOD console. This parameter is valid when Type is File. Notes: 1. Images less than 10M are recommended. 2. Image format values: jpeg, jpg, png.

Accessible file URL. This parameter is valid when Type is Url. Note: 1. Images less than 10 MB are recommended; 2. Supported image formats: jpeg, jpg, png.

Output file information for the AIGC video conversion task.

Task ID.

Task status. Valid values:

Error code. A non-zero error code is returned when a source error occurs. If 0 is returned, use the ErrCode of each specific task.

Extended error codes.

Enumeration values:

• RequestLimitExceeded: API concurrency limit exceeded.
• InvalidParameter.VoilationContent: The input prompt violates the Content Security Policy.
• InvalidParameterValue: Parameter value error.
• FailedOperation: Model tasks accumulate.
• InternalError: Internal error.
• InvalidParameter: Invalid parameters.
• InvalidParameter.MediaFormat: Invalid media format.
• ContentModerationFailed: Failed to pass content moderation.
• ResourceInsufficient: Insufficient resources.
• ModelGenerateFailed: Failed to generate model.
• ResourceNotFound: Resource not found.
• OperationCanceled: Operation canceled.
• TaskTimeout: Task timeout.

Error message.

Task progress, value ranges from 0 to 100.

Input information for the AIGC video generation task.

Output information of the AIGC video task.

Identifier for deduplication. If a request with the same identifier has been sent within the past seven days, an error is returned for the current request. The maximum length is 50 characters. If this is not specified or left empty, deduplication is not performed.

Source context. This is used to pass user request information. The task status change callback returns the value of this field. The maximum length is 1000 characters.

Model name.

Model version.

Input file information for an AIGC video generation task.

Fixed subject input information for the AIGC task.

Media file ID used as the tail frame to generate a video. This globally unique identifier of the file in VOD is assigned by the VOD backend after successful upload. You can obtain this field in the video upload completion event notification or the VOD console.

URL of the media file used as the last frame to generate video. Description:

1. Only GV, Kling, and Vidu models are supported. Other models are not currently supported. When ModelName is GV, if you specify this parameter, you must also specify FileInfos as the first frame of the video to be generated. When ModelName is Kling, ModelVersion is 2.1, and the output resolution Resolution is 1080P, you can specify this parameter. When ModelName is Vidu and ModelVersion is q2-pro or q2-turbo, you can specify this parameter.
2. The image size must be less than 5M.
3. 3. The value of the image format is: jpeg, jpg, png, webp.

Prompt for video generation. Supports up to 1000 characters. This parameter is required when FileInfos is empty.

Prompt to prevent the model from generating video. Supports up to 1000 characters.

Whether to optimize Prompt content automatically. When Enabled, the passed-in Prompt will be optimized automatically to enhance generation quality. Valid values:

Generation mode. Valid values:

Output result file output for AIGC image generation.

Region information of the input file. If the file url is a foreign address, selectable Oversea. Default: Mainland.

Scenario type. Values are as follows:

Random seed of the model.

Pass-through parameter.

Input video file type. Valid values:

File category. Valid values:

• Image: image;
• Video: video;
• Audio: audio;

Media file ID, the globally unique identifier of the file on VOD, assigned by the VOD backend after successful upload. You can obtain this field in the video upload completion event notification or the VOD console. This parameter is valid when Type is File. Description:

1. Images less than 10M are recommended.
2. Image format values: jpeg, jpg, png.

Accessible file URL. This parameter is valid when Type is Url.
Description:

1. An image less than 10M is recommended.
2. Image format values: jpeg, jpg, png.

Accessible file Base64. This parameter is valid when Type is Base64. Description:

1. The total size of all files must not exceed 7MB to avoid exceeding the 10MB upper limit of message size for cloud APIs after conversion to Base64.
2. Image format should be: jpeg, jpg, png, webp.
3. Video format should be: mp4, mov, avi.
4. Do not include prefixes such as data:image/jpeg;base64,.

Reference Type, applicable to GV, Kling, and PixVerse models.
Note:
When using the GV model, it can be used as a reference method. Available values: asset represents material, style represents style;
When using the Kling model and Category is Video, it can distinguish reference video types. feature represents a feature reference video, base represents the video to be edited;
When using the PixVerse model, it can be used for multi-image (subject) reference generation mode. Available values: subject represents subject, background represents background;

Usage: Vidu subject Id.
Vidu subject Id: prompt can be used via @subject Id. Valid when Category is Image.

Suitable for Vidu-q2 model.
When all images carry a subject Id, you can set a timbre Id for the subject. Valid when Category is Image. Timbre list: https://shengshu.feishu.cn/sheets/EgFvs6DShhiEBStmjzccr5gonOg

Whether to retain the original sound of the video. Valid when Category is Video. Values are as follows:

Used to distinguish whether the input image is for first (last) frame to video, image to video, or reference to video. Available values:

• FirstFrame: for the first frame of first (last) frame to video or image to video;
• Reference: for reference to video;
• LastFrame: for the last frame of first (last) frame to video;

Note: FirstFrame is selected by default.

Only effective for the multi-image (subject) reference mode of the PixVerse model. Assign a name to the image for more precise results. Usage: When this field value is "kitten", use @kitten in the Prompt to accurately describe the scenario. There must be a space after @Text, for example, @kitten run. The name referenced in the Prompt must be identical to this field.

Fixed subject Id.

• Kling subject required;
• Vidu subject selectable.

Fixed name.

• Vidu subject required. It can be used by adding [@name] in the prompt. For example, if the name is Xiao Ming, describe it as [@Xiao Ming] in the prompt.
• Kling subject selectable.

Valid only for Vidu. The voice type ID is used to determine the timbre of the sound in the video. If empty, the system will automatically recommend one.

Valid only for Vidu. Temporary subject images, up to 3 images
Note 1: Input image URLs are supported (ensure they are accessible);
Note 2: Images in png, jpeg, jpg, and webp formats are supported;
Note 3: Image pixels cannot be less than 128*128, and the ratio must be less than 1:4 or 4:1.

Valid only for Vidu. Temporary subject video, a maximum of 1 5-second video. Note 1: Only the viduq2-pro model supports the use of video subjects for reference; Note 2: Supports uploading up to 1 5-second video; Note 3: Video supports mp4, avi, mov formats; Note 4: Video pixel cannot be less than 128*128, and the ratio must be less than 1:4 or 4:1;

Output file information of the AIGC video generation task.

Task ID of the task whose task type is Procedure. If a task flow template (Procedure) is specified when initiating an AIGC video generation task, the task is initiated when the task flow template specifies one or more of MediaProcessTask, AiAnalysisTask, and AiRecognitionTask.

Usage information of AIGC video tasks.

Storage mode. Valid values:

Output filename, up to 64 characters. By default, the system generates the filename. Valid when StorageMode is Permanent.

Category ID, used to categorize and manage media. You can create a category and obtain the category ID through the Create Category API. Valid when StorageMode is Permanent.

Expiry date of the output file. The file will be deleted after this time. It never expires by default. Format according to the ISO 8601 standard. For details, see ISO date format description.

File type, for example, mp4 or flv.

Media file playback address.

File content. A return value is returned when UsageType is position_info.

Media file ID. Valid when StorageMode is Permanent.

Metadata of the output video.

Purpose type of the file.

Enumeration values:

• scene_url: 3D scene file. The FileUrl field has a return value.
• point_url: Point cloud file. The FileUrl field has a return value.
• mesh_url: Original mesh model file. The FileUrl field has a return value.
• mesh_simplified_url: Simplified mesh model file. The FileUrl field has a return value.
• position_info: Scene spatial position information. The FileContent field has a return value.
• image_url: Generated image. The FileUrl field has a return value.

Input the number of tokens.

Valid only when using the omni version of GV.

Number of tokens generated by thinking.

Valid only when using the omni version of GV.

Number of input images.

Valid only when using the H3 version of Hailuo.

Duration of the input video.

Unit: seconds.

Valid only when using the H3 version of Hailuo.

Output video duration.

Unit: seconds.

Valid only when using the H3 version of Hailuo.

Total input and output duration.

Default value: seconds.

Valid only when using Hailuo H3 version.

Voice full-text recognition task switch. Available values:

List of generated subtitle file formats. Leave empty or pass an empty array to indicate no subtitle file generation. Available values:

Media source language. Value range:

Specify subtitle name, length limited to 64 characters. This value will be used for player display. If left blank, VOD will auto generate it.
Note: This field is valid only when SubtitleFormats includes vtt.

Audio stream encoding format.
When the outer parameter Container is mp3, the valid value is:

Audio stream bitrate, in kbps. Value range: 0 and [26, 256]. When the value is 0, it means VOD automatically sets the bitrate.

Sampling rate of the audio stream. Available values:

Audio channel. Available values:

Specifies the retained audio tracks for output. All source tracks are retained by default.

This parameter is valid only when specified in OverrideParameter. In other cases, it does not take effect.