curl --request POST \
--url https://api.aiid.edu.kg/api/v1/services/aigc/video-generation/video-synthesis \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"model": "happyhorse-1.0",
"mode": "t2v",
"input": {
"prompt": "A white horse running through a neon city street, cinematic"
},
"parameters": {
"resolution": "720P",
"ratio": "16:9",
"duration": 5
}
}
'{
"id": "task_xxx",
"output": {
"task_id": "<string>",
"task_status": "<string>",
"video_url": "<string>"
},
"usage": {
"duration": "<string>",
"SR": "<string>"
}
}Create video (HappyHorse and Wan)
This series supports OpenAI video generation formats (see the link Video generation Sora compatible formats).
HappyHorse and the Wan video generation API, with public paths following the official DashScope protocol.
HappyHorse supports happyhorse-1.0, happyhorse-1.0-i2v, happyhorse-1.0-t2v, happyhorse-1.0-r2v, and happyhorse-1.0-video-edit; Wan supports wan-3.0, wan-3.0-480p, and wan-3.0-1080p. Among them, happyhorse-1.0 is an aggregated model and must be distinguished via mode into t2v, i2v, r2v, and video_edit. The three Wan models are fixed at 720P, 480P, and 1080P respectively, and support 2 to 30 seconds plus up to 10 image, 5 video, and 5 audio references.
The public API keeps the official endpoint format; channel-specific prefixes are only concatenated internally when the gateway requests a specific upstream channel. Clients do not need to pass X-DashScope-Async explicitly; it is handled by the gateway or the official channel adapter layer.
Provides the public request skeleton, task query structure, and recommended validation template for HappyHorse and Wan video generation formats; public paths follow the official DashScope endpoint format.
Supported models: happyhorse-1.0, happyhorse-1.0-t2v, happyhorse-1.0-i2v, happyhorse-1.0-r2v, happyhorse-1.0-video-edit, wan-3.0, wan-3.0-480p, wan-3.0-1080p
Unified request fields:
model(string, required): The external model name for HappyHorse or Wan. Wan supportswan-3.0,wan-3.0-480p, andwan-3.0-1080p; the HappyHorse aggregated modelhappyhorse-1.0must be used together withmode.mode(string, optional): Aggregated model mode; common values aret2v,i2v,r2v, andvideo_edit. You can omit this for fixed submodels.input.prompt(string, required): Prompt for video generation or editing.input.media(array[object], optional): Multimedia input. Supportsfirst_frame,last_frame,reference_image,reference_video, andreference_audio; Wan supports up to 10 images, 5 videos, and 5 audio files.parameters.duration(integer, optional): Generation duration, in seconds.parameters.resolution(string, optional): Output resolution. The three Wan models are fixed at720P,480P, and1080Prespectively; common values for HappyHorse are720Pand1080P.parameters.ratio(string, optional): Output aspect ratio, commonly used for text-to-video and reference-to-video.parameters.watermark(boolean, optional): Whether to add a watermark.parameters.seed(integer, optional): Random seed.
Common model/mode differences:
HappyHorse 聚合模型: Supports model happyhorse-1.0; the aggregated model uses mode to distinguish the final model.HappyHorse 文生视频: Supports model happyhorse-1.0-t2v; official-format text-to-video.HappyHorse 首帧图生视频: Supports model happyhorse-1.0-i2v; official-format first-frame image-to-video.HappyHorse 参考生视频: Supports model happyhorse-1.0-r2v; official-format multi-reference image-to-video.HappyHorse 视频编辑: Supports model happyhorse-1.0-video-edit; official-format video editing.Wan 3.0 720P: Supported model wan-3.0; Wan 3.0 official-format multimodal reference video generation.Wan 3.0 480P: Supported model wan-3.0-480p; Wan 3.0 official-format call fixed at 480P.Wan 3.0 1080P: Supported model wan-3.0-1080p; Wan 3.0 official-format call fixed at 1080P.
curl --request POST \
--url https://api.aiid.edu.kg/api/v1/services/aigc/video-generation/video-synthesis \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"model": "happyhorse-1.0",
"mode": "t2v",
"input": {
"prompt": "A white horse running through a neon city street, cinematic"
},
"parameters": {
"resolution": "720P",
"ratio": "16:9",
"duration": 5
}
}
'{
"id": "task_xxx",
"output": {
"task_id": "<string>",
"task_status": "<string>",
"video_url": "<string>"
},
"usage": {
"duration": "<string>",
"SR": "<string>"
}
}Authorizations
Use Bearer Token authentication.
Format: Authorization: Bearer sk-xxxxxx
Body
HappyHorse or Wan public model name. Wan supports wan-3.0, wan-3.0-480p, and wan-3.0-1080p; the HappyHorse aggregated model happyhorse-1.0 must be used with mode.
happyhorse-1.0, happyhorse-1.0-t2v, happyhorse-1.0-i2v, happyhorse-1.0-r2v, happyhorse-1.0-video-edit, wan-3.0, wan-3.0-480p, wan-3.0-1080p "happyhorse-1.0-t2v"
Hide child attributes
Hide child attributes
Prompt for video generation or editing.
"A white horse running through a neon city street, cinematic"
Multimedia input. Supports first_frame, last_frame, reference_image, reference_video, and reference_audio; Wan supports up to 10 images, 5 videos, and 5 audio files.
Aggregated model mode. Common values are t2v, i2v, r2v, and video_edit. The fixed submodel is optional.
t2v, i2v, r2v, video_edit Hide child attributes
Hide child attributes
Output resolution. The three Wan models are fixed at 720P, 480P, and 1080P respectively; the common values for HappyHorse are 720P and 1080P.
480P, 720P, 1080P "720P"
Output aspect ratio, commonly used for text-to-video and reference-to-video.
"16:9"
Generation duration in seconds.
5
Whether to add a watermark.
Random seed.
Response
Task created successfully
"task_xxx"