HappyHorse Video API
Use a stable REST API to cover text-to-video, first-frame-to-video, reference-image-to-video, and video editing. Suitable for integrating creative storyboards, product materials, character references, and post-editing processes into your own products.
Key Actions for Video Production
No need to stitch processes between multiple vendor interfaces, HappyHorse consolidates common video generation and editing modes into a unified request structure.
Text-to-Video
Input shot language, scenes, and actions to quickly generate short video prototypes.
First-Frame-to-Video
Control composition with a fixed first frame, allowing static visual materials to come to life.
Reference-Image-to-Video
Constrain characters, props, styles, and visual elements through multiple reference images.
Video Editing
Perform style transfer, outfit changes, and local replacements based on input videos and reference images.
curl -X POST 'https://api.acedata.cloud/happyhorse/videos' \
-H 'authorization: Bearer YOUR_API_KEY' \
-H 'accept: application/json' \
-H 'content-type: application/json' \
-d '{
"action": "image_to_video",
"model": "happyhorse-1.1-i2v",
"image_url": "https://cdn.acedata.cloud/b1c82e4937.png",
"prompt": "A cinematic white horse lifts its head, the mane moves gently in the sunrise wind, slow camera push in, warm film lighting",
"resolution": "720P",
"duration": 5,
"async": true
}'Submit a video task
Use async to immediately return the task ID, and then query the results through `/happyhorse/tasks`; if a `callback_url` is provided, the final result will be pushed to your service after the task is completed.
Select action
Text-to-video, image-to-video, reference images, or editing all use the same entry point.
Provide materials
Images and videos must be publicly accessible URLs, it is recommended to use Ace Data Cloud CDN.
Query results
Get the final CDN video URL, duration, and resolution based on task_id.
Three types of scenarios suitable for integration into real products
From marketing materials to content tools, HappyHorse is more suitable as a stable video capability module rather than a one-time toy.
Brand short film prototype
Quickly generate product, character, and scene short films using text and the first frame to validate visual direction.
Consistency of characters and props
Lock in characters, costumes, props, and visual styles through multiple reference images.
Video post-editing
Perform style transfer, local replacement, and audio retention strategies on existing clips.
Three steps to launch video capabilities
Continuing with Ace Data Cloud's unified authentication, billing, task, and callback model, the integration path is clear and controllable.
Apply for Token
Create an application in the console to obtain a unified Bearer Token.
Submit task
Pass in prompt, image_url, image_urls, or video_url according to action.
Fetch results
Obtain downloadable CDN video URL through task API or callback.
Why Use a Unified API
Consolidate model calls, task polling, CDN storage, error classification, and billing records on the platform side.
| Capability | Ace Data Cloud HappyHorse | Some Others |
|---|---|---|
| Multiple video actions | Unified `/happyhorse/videos` | Multiple interfaces to stitch together |
| Result files | Automatic CDN storage | Common temporary URLs expire |
| Billing | Records based on actual seconds | Requires self-built usage ledger |
| Error handling | Structured error codes | Upstream error formats are unstable |
Select Model by Task
1.1 Models are primarily used for generating tasks; video editing currently uses 1.0 video-edit.
happyhorse-1.1-t2v
Default text-to-video model, suitable for generating short films from scripts and shot descriptions.
happyhorse-1.1-i2v
Controls composition and visual tone with the first frame image, suitable for product images and character image animation.
happyhorse-1.1-r2v
Constrains characters and props with multiple reference images, suitable for scenes with higher consistency requirements.
happyhorse-1.0-video-edit
Edits based on the input video and can choose to retain the original audio.
Clear Per-Second Pricing
Failed tasks are not billed; video editing is billed based on the total input and output duration returned by upstream.
720P Generation
- Suitable for rapid prototyping
- Supports t2v / i2v / r2v
- 3–15 seconds output
1080P Generation
- Clearer delivered materials
- Suitable for commercial previews
- Unified task queries
Video Editing
- Input + output duration
- Supports reference images
- Can retain original audio
Frequently Asked Questions
The parameters, materials, and billing issues that are most concerning before going live.
What input materials does HappyHorse support?
Text-to-video only requires a prompt; first-frame image-to-video requires image_url; reference image-to-video requires 1–9 image_urls; video editing requires video_url, and optionally 0–5 reference images.
Will the returned video URL expire?
Ace Data Cloud will transfer upstream results to the platform CDN, and the accessible CDN URL is used in the documentation and API responses.
How to handle long generation times?
It is recommended to use async or callback_url. The interface will first return task_id, and you can actively query through `/happyhorse/tasks` or wait for a callback.
Will failed tasks incur charges?
No. Only tasks that are successfully completed and recorded will be billed based on duration and resolution.
Related Video Capabilities
If you need different model ecosystems or more segmented video capabilities, you can combine these public services.
Integrate HappyHorse into your video product
Use unified Token, unified billing, and unified task queries to stably integrate video generation and editing capabilities into your application.