Skip to content

Model Description

Capability Map of Each Model

happyhorse-1.0-t2v happyhorse-1.0-i2v happyhorse-1.0-r2v happyhorse-1.0-video-edit
Text-to-Video ✅
Image-to-Video ✅
Reference student video ✅
Edit Video ✅

Function Interface Details

Video URLs are only retained for 24 hours and will be automatically deleted after the timeout. Please save the generated video in a timely manner.

Create a video generation task (unified interface for text-to-video, image-to-video, reference-based video generation, and video editing)

Request URL

POST https://{新平台域名}/v1/ai/{endpointPath}/happyhorse/v1/video-synthesis

Request Body Parameters

Parameter Name Field Type Required or Not Default Value Description
model
 string is - 模型名称，固定值：happyhorse-1.0-t2v / happyhorse-1.0-i2v / happyhorse-1.0-r2v / happyhorse-1.0-video-edit
input object is - Input Object
input.prompt
 string No - Text prompt. Used to describe the expected generated video content. Supports input in any language, with a length not exceeding 5000 non-Chinese characters or 2500 Chinese characters; any excess will be automatically truncated.
input.media array No - Input media list. Used to specify the video to be edited and the reference image
Used in scenarios such as image-to-video, reference-to-video, and video editing,for details, see the description below the table
input.media[].type
 string No - first_frame / reference_image /video/ reference_image
See the instructions below the table
input.media[].url string No - Media Creatives URL
See the instructions below the table
parameters object No - Generation Parameters
parameters.resolution
 string No 1080P Resolution levels for generated videos
720P / 1080P
parameters.ratio string No 16:9 16:9 / 9:16 / 1:1 / 4:3 / 3:4
Only used in text-to-video/reference-to-video
parameters.duration
 integer No 5 Duration (seconds), an integer between 3 and 15
Only used in text-to-video/image-to-video/reference-to-video
parameters.watermark
 boolean No true Whether to add watermark
The watermark is located in the bottom right corner of the video, with the text fixed as Happy Horse.
parameters.audio_setting
 string
 No auto auto: The model controls itself
origin: Retain the original audio of the input video
For video editing use only
parameters.seed
 integer No - Random seed [0, 2147483647]
Please note that due to the probabilistic nature of model generation, even when using the same random seed, it cannot be guaranteed that the results of each generation will be completely identical.

Text-to-Video:

  • input.prompt is required

Image-to-Video:

  • input.prompt is optional, input.media.type is required, and input.media.url is required

  • The optional values for the input.media.type field are: first_frame, Creative restrictions: There must be exactly 1 first frame image.

  • input.media.url

First frame URL. Supports HTTP or HTTPS protocol.

Image Restrictions:

  • Format: JPEG, JPG, PNG, WEBP.

  • Resolution: Width and height are not less than 300 pixels.

  • Aspect ratio: 1:2.5 to 2.5:1.

  • File size: no more than 10MB.

Reference student video:

  • input.prompt is required, input.media.type is required, input.media.url is required

  • Reference Indicators: In the prompt, usecharacter1,character2to identify the reference images at corresponding positions in the media array, with the order consistent with that of the media array.

  • Each element of the media array is a media object, containing the type and url fields.

  • Define the order of role references in the prompt according to the array order.

  • The first reference_image in the array corresponds to character1 , the second corresponds to character2 , and so on.

  • input.media.type is fixed toreference_image, number of reference images: 1 to 9.

  • input.media.url supports HTTP or HTTPS protocols.

  • Format: JPEG, JPG, PNG, WEBP.

  • Resolution: The short side should not be less than 400 pixels, and clear images above 720P are recommended. Avoid uploading images that are too small, blurry, or over-compressed, as they may affect the result.

  • File size: no more than 10MB.

Video Editing:

  • input.prompt is required, input.media.type is required, input.media.url is required

  • input.media.type options:video,reference_image, there must be exactly 1 input.media of type video, and 0-5 of type reference_image

  • input.media.url (type=video)

  • Format: MP4, MOV (H.264 encoding recommended).

  • Duration: 3 to 60 seconds.

  • Resolution: The long side should not exceed 2160 pixels, and the short side should not be less than 320 pixels.

  • Aspect ratio: 1:2.5 to 2.5:1.

  • File size: no more than 100MB.

  • Frame rate: greater than 8fps.

Description

Output video duration: 3 - 15 seconds.

  • When the input video is no longer than 15 seconds, the duration of the output video remains consistent with that of the input video.

  • When the input video exceeds 15 seconds, the system will automatically capture the first 15 seconds from the beginning as the valid segment, so the maximum output duration is 15 seconds.

  • input.media.url (type=reference_image)

  • The URL of the reference image must be a publicly accessible URL.

  • Supports HTTP or HTTPS protocol.

  • Format: JPEG, JPG, PNG, WEBP.

  • Resolution: Width and height dimensions are not less than 300 pixels.

  • Aspect ratio: 1:2.5 to 2.5:1.

  • File size: no more than 10MB

Response Parameter

Parameter Name Field Type Description
output object Task Output Information
output.task_id string Task ID (for querying results)
output.task_status
 string PENDING / RUNNING / SUCCEEDED / FAILED / CANCELED / UNKNOWN
request_id string Request Unique Identifier
code string Failure error code, this parameter will not be returned when the request is successful
message string Failure error message, this parameter will not be returned when the request is successful

Request Example

curl --location 'http://{新平台域名}/v1/ai/{endpoint}/happyhorse/v1/video-synthesis' \
--header 'Authorization: Bearer {token}' \
--header 'Content-Type: application/json' \
--data '{
    "model": "happyhorse-1.0-t2v",
    "input": {
        "prompt": "A miniature city built of cardboard and bottle caps comes to life at night. A cardboard train moves slowly through it, dotted with small lights that illuminate the way."
    },
    "parameters": {
        "resolution": "720P",
        "ratio": "16:9",
        "duration": 5
    }
}'

Response Example

{
  "output": {
        "task_status": "PENDING",
        "task_id": "0385dc79-5ff8-4d82-bcb6-xxxxxx"
    },
    "request_id": "4909100c-7b5a-9f92-bfe5-xxxxxx"
}

Query task results

Request URL

GET https://{新平台域名}/v1/ai/{endpointPath}/happyhorse/v1/tasks/{taskId}

Path Parameter

Parameter Name Field Type Required or Not Default Value Description
taskId string is - output.task_id returned by creating a task

Response Parameters

Parameter Name Field Type Description
output object Task Output Information
output.task_id string Task status, query validity period is 24 hours.
output.task_status string Task Status
output.submit_time string Task submission time. The time zone is UTC+8, and the format is YYYY-MM-DD HH:mm:ss.SSS.
output.scheduled_time
 string Task execution time. The time zone is UTC+8, and the format is YYYY-MM-DD HH:mm:ss.SSS.
output.end_time string Task completion time. The time zone is UTC+8, and the format is YYYY-MM-DD HH:mm:ss.SSS.
output.video_url
 string Video URL. Only returned when task_status is SUCCEEDED.
The link is valid for 24 hours, and the video can be downloaded via this URL. The video has a frame rate of 24fps and is in MP4 format (H.264 encoding).
output.orig_prompt string The original input prompt, corresponding to the request parameter prompt.
output.code string Failure error code, this parameter will not be returned when the request is successful
output.message string Failure error message, this parameter will not be returned when the request is successful
usage object Dosage (duration, SR, ratio, etc.)
usage.input_video_duration integer Duration of the input video, in seconds.
usage.output_video_duration integer Duration of the output video, in seconds.
usage.duration integer Total video duration, used for billing.
usage.SR integer Resolution level of the output video.
usage.video_count integer Number of output videos. Fixed at 1
request_id string Request Unique Identifier

Request Example

curl --location 'http://{新平台域名}/v1/ai/{endpoint}/happyhorse/v1/tasks/{taskId}' \
--header 'Authorization: Bearer {token}' \
--header 'Content-Type: application/json'

Response Example

{
    "request_id": "99243b47-ec5f-9413-9993-xxxxxx",
    "output": {
        "task_id": "4673458e-28be-4a05-bf2a-xxxxxx",
        "task_status": "SUCCEEDED",
        "submit_time": "2026-04-20 17:55:17.075",
        "scheduled_time": "2026-04-20 17:55:17.129",
        "end_time": "2026-04-20 17:56:36.658",
        "orig_prompt": "A miniature city built of cardboard and bottle caps comes to life at night. A cardboard train moves slowly through it, dotted with small lights that illuminate the way.",
        "video_url": "https://dashscope-result.oss-cn-beijing.aliyuncs.com/xxx.mp4?Expires=xxx"
    },
    "usage": {
        "duration": 5,
        "input_video_duration": 0,
        "output_video_duration": 5,
        "video_count": 1,
        "SR": 720,
        "ratio": "16:9"
    }
}