OpenAI-style auth
Use bearer API keys in the Authorization header. Keys are stored as hashes and shown only once.
Seedance 2.0 video generation API
Build with Seedance 2.0 cinematic video generation.
Seedance API makes Seedance 2.0 available through a simple developer workflow for text-to-video, image-to-video, async task polling, credits, API keys, and full request logs.
Text to video Image to video Async polling
Official Seedance 2.0 showcasePrompt, image URL, duration, resolution, task status
Task IDs and polling
Every generation returns a local video task ID for status checks and result retrieval.
Log-first debugging
Inspect client requests, normalized payloads, upstream responses, status codes, and latency.
What
What is Seedance API?
Seedance API is a developer access layer for the Seedance video generation model family. The homepage focuses on Seedance 2.0 because it is the clearest default for teams that want cinematic text-to-video, image-to-video, and reference-guided video generation in a production product.
Instead of treating video generation as a one-off demo, Seedance API packages the pieces developers need to ship a real workflow: bearer API keys, credit-based usage, async video tasks, polling, request logs, upstream response visibility, and a dashboard for reviewing what happened after a request was sent.
The core generation flow is intentionally straightforward. Your product sends a prompt, chooses a model such as seedance-2.0, passes duration and resolution controls, optionally includes a reference image URL, and receives a video task ID. Your app then polls that task until the generated video is ready or until a failure can be inspected in the logs.
Where
Where Seedance API fits in your product stack.
Seedance API is designed to sit between your application and the Seedance video generation workflow. It belongs in the part of your product where prompts, images, user intent, credits, generation status, and generated videos need to become a reliable user experience.
Backend API integration
Call Seedance API from your server, worker, or API route. Keep bearer keys away from the browser, send structured JSON payloads from your backend, and return task status or generated video URLs to your own frontend.
Product dashboard operations
Use the dashboard as the operational layer for credits, API keys, logs, and video task history. This is where developers can inspect what happened after a generation request leaves their application.
Creative automation pipelines
Connect prompts, reference images, duration, resolution, and polling into repeatable workflows. Seedance API can sit behind campaign tools, ecommerce automation, internal batch generation, or prototype studios.
How
Create a Seedance video task with one POST request.
Use a bearer API key, send a prompt with generation controls, then poll the task ID returned by the API. The shape is intentionally familiar for teams already building with modern AI APIs, but the surrounding dashboard makes the workflow easier to operate after launch.
cURLPythonJavaScriptCopy
1curl https://api.seedanceapi.app/v1/videos/generations \2 -X POST \3 -H "Authorization: Bearer sk_live_xxx" \4 -H "Content-Type: application/json" \5 -d '{6 "model": "seedance-2.0",7 "prompt": "A cinematic product reveal with smooth camera motion",8 "image_url": "https://example.com/reference.png",9 "duration": 5,10 "resolution": "720p"11 '}'
ResponsePoll task
1{2 "id": "vid_01jx9w6n8e6r7x9m2p4q",3 "object": "video.generation",4 "status": "running",5 "created_at": 17823600006}
Seedance 2.0 capabilities
Expose the creative controls users expect from modern video models.
The public API surface is built for apps that need text prompts, reference images, cinematic motion, duration and resolution controls, and a clean task lifecycle.
Text to video
Turn cinematic prompts, brand concepts, product shots, and scene directions into video generation tasks. Seedance 2.0 is especially useful when your product needs users to describe camera movement, scene pacing, subject action, visual mood, and output length in natural language.
Image to video
Start from a reference image URL and generate motion while keeping the subject, style, and scene direction close to the input. This is the workflow many apps need for product photos, character concepts, fashion images, food shots, thumbnails, and campaign visuals.
Reference-guided output
Build workflows around character references, product references, shot composition, camera motion, and visual consistency. Reference-guided generation gives teams a bridge between open-ended prompt creativity and controlled brand or product assets.
Cinematic motion
Expose prompt controls for camera moves, close-ups, wide shots, pacing, lighting, and realistic movement. The page should attract users who search for a cinematic AI video API, not only users who already know the Seedance model name.
Flexible formats
Pass duration, resolution, frame rate, and source media fields through one stable JSON request shape. A stable request surface helps product teams test 5-second previews, higher-resolution outputs, and app-specific generation presets without changing their whole integration.
Async tasks
Return a task ID immediately, keep upstream responses traceable, and poll status without blocking your app. Async design matters because video generation can take longer than a normal web request, especially when users submit richer prompts or reference images.
Official examples
Official Seedance 2.0 examples for API builders.
These examples use media from the Seedance 2.0 official product demonstrations, but the presentation is adapted for this API product site: each example explains the inputs, the generated output, and how the capability maps to a developer workflow.
Official Seedance 2.0 example
Multimodal reference-based video generation.
with fine-grained feature preservation
Use video, image, and audio references together when a product workflow needs stronger control than prompt-only generation. This official Seedance 2.0 example shows how reference motion, character imagery, scene direction, and audio cues can guide the final video while preserving recognizable details.
Video referencesAudio referencesImage references
Reference
Output
Prompt: Generate a pixel-style fight scene from a motion reference, character images, a background image, and an audio reference while keeping the action and sound direction aligned.
Official Seedance 2.0 example
Precise, targeted video editing.
subject replacement, object-level editing, and inpainting
Use targeted editing when the original motion should stay stable but one subject, object, or region needs to change. This official example shows the kind of API workflow a product can expose when users need controlled video edits rather than a full regeneration from scratch.
Subject replacementObject-level editingInpainting
Video edit
Output
Prompt: Replace the product subject in the gift-box video with the referenced cream jar while preserving camera motion, lighting, and the surrounding scene.
Why
Why use a Seedance API layer instead of only a model demo?
A model demo proves what the model can do. A product API layer helps your team turn that capability into a repeatable workflow with authentication, usage control, status tracking, and operational visibility.
A safer way to expose generation
Most products should not expose upstream provider keys directly to the browser. Seedance API gives your product a server-side integration pattern with bearer keys, hashed storage, and task records, so your frontend can stay focused on prompts, previews, and user experience.
Visibility after the request leaves your app
Video generation failures can happen during validation, upstream task creation, polling, or result retrieval. A product-ready API layer should keep the normalized request, upstream response, status code, and latency visible so developers can debug real customer workflows.
Stable controls for product UX
A user interface needs consistent controls for model, prompt, image URL, duration, resolution, and frame rate. Seedance API keeps those controls in a stable JSON request shape, which makes it easier to build presets, templates, and repeatable generation flows.
Async by default for video workloads
AI video generation does not behave like a short text completion. Users expect progress, retries, and result history. Returning a task ID first lets your app show status, keep the interface responsive, and handle longer-running Seedance 2.0 generations predictably.
How it works
From checkout to generated video without building billing or key management yourself.
Seedance API combines the generation endpoint with the operational pieces needed for a real video generation product. The flow is short enough for a quick prototype, but explicit enough for teams that need auditability, predictable task state, and a place to inspect failures.
1
Create an account and buy credits
Credits give your API keys permission to create video tasks. This keeps the first integration simple: purchase a pack, create a key, and start testing generation requests without building a billing system first.
2
Generate a bearer API key in the dashboard
Use the key from your backend in the Authorization header. Keys are shown once and stored as hashes, so your team can build around a familiar bearer-token pattern while keeping credentials controlled.
3
Send a JSON generation request
Include a prompt, model, duration, resolution, optional image URL, and frame-rate settings. The endpoint returns a local task ID so your product can track the generation independently from the upstream provider response.
4
Poll status and display the result
Poll the task endpoint until the status is succeeded or failed. When the task succeeds, your app can read the result, display the generated video, save it to a project, or pass it into another creative workflow.
Who
Who uses Seedance API?
The best users for Seedance API are teams that need AI video generation inside a product, not just a model demo. These teams usually need stable API calls, task status, credit visibility, and enough logging to understand what happened when a generation succeeds or fails.
Developers building AI video apps
Use Seedance API when your product needs a video generation backend but your team does not want to build authentication, key storage, credit tracking, request logging, and task polling from scratch.
SaaS teams adding video features
Add text-to-video or image-to-video workflows to an existing app. Your users can submit prompts and reference images while your backend handles API calls, task status, and result retrieval.
Ecommerce and product teams
Create product video experiments from product photos, reference images, and campaign prompts. This is useful for testing motion concepts before investing in manual editing or production work.
Creative agencies and ad teams
Prototype campaign concepts, social video directions, mood boards, and story variations. Seedance 2.0 can help teams move from written creative direction to video output faster.
Internal AI platform teams
Evaluate Seedance workflows inside controlled internal tools. Logs, task history, and credit visibility help teams understand usage before rolling video generation into a public product.
Media workflow builders
Use the API as one step in a larger workflow that may include prompt templates, asset uploads, moderation, review queues, downloads, analytics, and customer-facing creative controls.
Build examples
What you can build with Seedance 2.0 and Seedance API.
The same API surface can support several product directions. The strongest SEO value comes from explaining concrete workflows, not only listing model features.
Text-to-video generator
Build an app where users describe a scene, choose duration and resolution, submit the prompt, and receive a generated video after polling. This is the most direct Seedance 2.0 workflow for creators, marketers, and product teams.
Image-to-video product tool
Let users upload or provide a product image, then animate it into a short video concept. Ecommerce teams can use this workflow to test product motion, launch visuals, social ads, and landing-page creative before production.
Reference-based creative workflow
Combine prompts with reference images for characters, products, environments, or visual style. This helps teams move beyond generic AI video output and gives them a repeatable way to keep generated videos closer to an intended brand direction.
Internal batch testing dashboard
Create an internal dashboard where teams test many prompts, compare task outcomes, inspect logs, and refine generation templates. This is useful before making AI video generation available to customers inside a public-facing product.
Workflow choice
How to choose the right Seedance 2.0 generation workflow.
The best workflow depends on where the user's creative intent begins. Some products start from a written prompt, some start from an image, and some need reference control so generated videos stay closer to a product, character, or campaign direction.
Use text-to-video when the idea starts as a scene
Choose text-to-video when your user knows the desired story, camera movement, mood, or action but does not have a fixed source image. This works well for concept videos, scene exploration, ad ideas, visual storytelling, and early creative directions where language is the main input.
Use image-to-video when the source asset matters
Choose image-to-video when the starting point is a product photo, character image, fashion shot, food image, or visual reference that should remain recognizable. This workflow is often better for ecommerce, brand teams, and apps where users expect motion without losing the source subject.
Use reference-guided workflows when consistency matters
Choose reference-guided generation when your product needs stronger control over style, subject identity, composition, or brand direction. Reference inputs help teams move from generic AI video output toward repeatable creative systems that can be tested, reviewed, and improved over time.
FAQ
Seedance API questions developers ask first.
These answers reinforce the core search intent behind the page: what Seedance API is, where it fits, who should use it, and how the integration works from the first request to the final generated video.
What is Seedance API?
Seedance API is a developer access layer for the Seedance video generation model family. It focuses on making Seedance 2.0 practical for product teams that need text-to-video, image-to-video, reference-guided generation, task polling, API keys, credits, and request logs in one workflow.
Is Seedance API focused on Seedance 2.0?
Yes. The Seedance API product is positioned around the Seedance family, while the homepage and quickstart flow highlight Seedance 2.0 as the primary generation workflow. That gives developers a clear default model while leaving room for future Seedance model options.
Where should I call Seedance API from?
Most teams should call Seedance API from a backend service, worker, or server-side route rather than exposing API keys in the browser. Your frontend can collect prompts and reference images, then your backend can create tasks, poll results, and return generated video URLs to the user interface.
Who is Seedance API for?
Seedance API is for developers, SaaS founders, product teams, creative automation teams, ecommerce teams, and agencies that want to add AI video generation without building billing, key management, task storage, and debug logs from scratch.
Does Seedance API support image-to-video workflows?
Yes. The create request can include an image URL alongside the prompt, model, duration, resolution, and frame-rate fields. That makes it suitable for product shots, character references, scene references, and other workflows where a static image should guide generated motion.
How does async task polling work?
A create request returns a local video task ID immediately. Your application polls the task endpoint until the status changes from running to succeeded or failed. When the generation succeeds, the response can expose the generated video result for your application to display, save, or pass into another workflow.
How do credits work?
Credits are attached to API keys and are reserved when a create request is accepted. The current product copy keeps the pricing model simple for developers: buy credits, create a key, call the endpoint, and use dashboard logs to understand usage and failures.
How do I debug failed generations?
The dashboard keeps request logs, normalized payloads, upstream status codes, upstream response bodies, task status, and latency. That means a failed generation can be traced through authentication, validation, upstream creation, polling, and result retrieval instead of becoming a silent black box.