Video API

One hosted video API, three public model tracks

Use the shared API workflow for task creation, polling, and key management, then route traffic into the video model family that matches your business goal.

POST /api/v1/jobs/createTaskGET /api/v1/jobs/recordInfoShared API keys

Model tracks

Public video routing should stay narrow, legible, and intent-led

Every page below can share the same underlying job lifecycle while still presenting a clean model-specific promise to the buyer.

Seedance 2

The current production-facing Seedance API surface for text-to-video and image-to-video workflows.

View details

Seedance 2.5

A release-aware Seedance route that can absorb launch demand while the live generation path stays grounded in today’s backend.

View details

Happy Horse

A premium video option with native audio, reference-aware generation, and strong product-storytelling fit.

View details

Endpoints

The public video API should feel small enough to learn in one pass

Keep the external contract simple: create a task, poll the task, and let the model selection happen through a small set of explicit model ids.

POST/api/v1/jobs/createTask

Create a video generation task

Use one endpoint for text-to-video and image-to-video tasks, then let the model page explain which track the request belongs to.

{
  "model": "seedance-2-0",
  "inputs": {
    "prompt": "A premium product reveal in a moody studio",
    "duration": "5s",
    "aspectRatio": "16:9",
    "resolution": "720p"
  }
}

GET/api/v1/jobs/recordInfo?taskId=task_xxx

Poll task status

Use one polling contract across every public video track, then display the final output URL when the job completes.

{
  "code": 200,
  "message": "success",
  "data": {
    "taskId": "task_xxx",
    "status": "success",
    "output": [
      {
        "url": "https://cdn.example.com/video.mp4"
      }
    ]
  }
}

Workflow

The public video API flow should feel identical across model pages

Authenticate once

Issue an API key, attach a Bearer token, and keep one auth shape across every public video track.

Create the task

Send a request to POST /api/v1/jobs/createTask with the target model and your prompt or source asset URLs.

Poll the job

Use GET /api/v1/jobs/recordInfo?taskId=... to track progress until the final output becomes available.

Route by model intent

Use Seedance 2 for the live surface, Seedance 2.5 for launch intent, and Happy Horse for premium reference-heavy video jobs.

Routing guide

When each video API page is the right destination

Seedance 2

Best when you need the current production-facing video API page with honest pricing, onboarding, and working endpoints.

Seedance 2.5

Best when growth needs a release-day landing page without pretending that the backend is already switched over.

Happy Horse

Best when you want richer premium positioning around native audio, reference-aware generation, and product storytelling.

Back to API docs Get API key

Model comparison

Which public video page should carry which kind of buyer

Seedance 2

Current production API page

Best for the live public route with stable onboarding and direct API framing.

Open page

Seedance 2.5

Launch-intent capture

Best for forward-looking demand capture while the actual execution path stays grounded in the current backend.

Open page

Happy Horse

Premium reference-heavy video

Best for native audio, richer reference behavior, and stronger premium storytelling positioning.

Open page

Billing logic

Pricing behavior the page can safely promise

Credits are checked and frozen before task execution.

Billing stays tied to the selected model and requested runtime.

Polling and ownership checks stay consistent across the whole public video API surface.

FAQ

The quick clarifications technical buyers will ask

Do all video models use the same API contract?

At the public surface level, yes. The createTask and recordInfo flow stays consistent while model pages explain positioning and workflow differences.

Why split Video API from model-specific pages?

The Video API page explains the shared contract. The model-specific pages absorb release intent, campaign traffic, and positioning without cluttering the base contract.

Where should customers start?

Start at Video API if they need the shared workflow. Send them straight to Seedance 2, Seedance 2.5, or Happy Horse if they already know the model family they want.

開始使用I2V創作

加入全球數百萬使用I2V業界領先AI影片工具的創作者行列。將你的個人或專業點子轉化為精致且可直接使用的影片片段——無需製作團隊或複雜的後期剪輯。

立即開始創作查看範例加入Discord

The public video API flow should feel identical across model pages

Authenticate once

Issue an API key, attach a Bearer token, and keep one auth shape across every public video track.

Create the task

Send a request to POST /api/v1/jobs/createTask with the target model and your prompt or source asset URLs.

Poll the job

Use GET /api/v1/jobs/recordInfo?taskId=... to track progress until the final output becomes available.

Route by model intent

Use Seedance 2 for the live surface, Seedance 2.5 for launch intent, and Happy Horse for premium reference-heavy video jobs.