AI Model APIVideos
HappyHorse Video API
How to call HappyHorse video models through New API.
HappyHorse video models can be called through two API styles:
- HappyHorse native style:
POST /happyhorse/api/generate,GET /happyhorse/api/status/{task_id} - New API generic video style:
POST /v1/video/generations,GET /v1/video/generations/{task_id}
Both styles submit tasks to the HappyHorse upstream channel. The difference is the request and response shape:
/happyhorse/api/*keeps the HappyHorse-style request and response./v1/video/generations*keeps the New API generic video task format.
Supported Models
| Model | Capability | Description |
|---|---|---|
happyhorse-1.0-t2v | Text to video | Generate a video from a text prompt |
happyhorse-1.0-i2v | Image to video | Generate a video from one first-frame image |
happyhorse-1.0-r2v | Reference to video | Generate a video from 1 to 9 reference images |
happyhorse-1.0-video-edit | Video edit | Edit one input video, with optional reference images |
Authentication
All requests must include a New API token:
Authorization: Bearer sk-xxxx
Content-Type: application/jsonParameter Rules
| Parameter | Rule |
|---|---|
prompt | Required. The generation or editing prompt |
duration | Supported by T2V, I2V, and R2V. If explicitly provided, it must be 3 to 15 seconds |
resolution | Supports 720P and 1080P |
ratio | Only supported by T2V and R2V. Supports 16:9, 9:16, 1:1, 4:3, 3:4 |
| Image input | Supports HTTP/HTTPS URL or image base64 data URL, for example data:image/png;base64,... |
| Video input | Supports HTTP/HTTPS URL only |
Media Input Rules
Image media types:
first_frame: the I2V first-frame image. Exactly 1 image is required.reference_image: reference images for R2V or Video Edit.
Video media type:
video: the Video Edit input video. Exactly 1 HTTP/HTTPS URL is required.
Image base64 must use a data URL prefix:
data:image/png;base64,iVBORw0KGgoAAAANSUhEUg...Plain base64 strings, non-image MIME types, and empty base64 content are rejected. Video base64 is not supported.
Billing
HappyHorse uses the New API task billing flow. Quota is pre-consumed when a task is submitted, then settled after completion based on upstream usage.
- T2V, I2V, R2V: billed by output video duration.
- Video Edit: billed by input video duration plus output video duration.
- The final price depends on model price and group ratio configured in New API.
Next Steps
How is this guide?