In the evolving landscape of AI-powered video generation, Apiframe has stepped forward with its Seedance 2.0 API, providing a consolidated, highly flexible interface for creative developers and ad-tech teams. This API layers complex multimedia synthesis into an elegant one-endpoint experience for generating videos from text, images, and even reference videos. Among its pioneering features are multimodal references with defined roles, native synchronized audio within a single generation pass, and director-style camera movements controlled directly via prompt language.
If you’ve previously explored tools from ByteDance and CapCut, you’ll appreciate how Apiframe’s approach advances seamless multimedia creation at scale. But if you’re asking, “Where are the fields listed in the Seedance 2.0 API docs?” or need clarity on the seedanceParams input schema, this post serves as a detailed guide.
Understanding the Single Endpoint Paradigm
Traditional video generation APIs tend to fragment capabilities — separate endpoints for text-to-video, image-to-video, or motion-transfer tasks. Seedance 2.0 breaks this mold by providing a single, unified API endpoint:
POST https://api.apiframe.ai/v2/videos/generateThis endpoint accepts a comprehensive JSON payload that captures your input modality, references, style choices, and audio parameters all in one go. Whether you want to produce a short text-driven clip, stylize an image sequence, or integrate motion from a reference video, it’s all possible with a single request.
Why this matters
- Simplified integration: You only need to manage one endpoint and track one job ID. Consistent billing: You are billed per second of final video output, irrespective of input type. Multimodal flexibility: Mix and match text, images, videos, and audio references with clear role assignments.
Where Are the Fields Listed for Seedance 2.0?
Many developers wrestling with the Seedance 2.0 API first question is: “Where exactly are the input fields documented?” The Apiframe docs present the seedanceParams seed parameter reproducible schema within the POST video generation endpoint. These fields are detailed both in the JSON schema definitions and in prose examples, but they’re sometimes nestled inside larger API references, making them easy to overlook.
To save time, here’s a high-level breakdown of the key fields you’ll find under seedanceParams:
The seedanceParams schema is flexible but always benefits from sanity-checking defaults. For example, the resolution will default to 1280x720 if unspecified, and generate_audio defaults off unless explicitly enabled.
Multimodal References with Roles: Style, Motion, and Sound
One of Seedance’s most compelling capabilities lies in its intelligent handling of references. Instead of clunky disparate parameters, you assign each input media a role, clarifying its purpose for the generation engine:
- style: Defines the visual style, coloring, or aesthetic of the output. motion: Maps the reference video’s motion vectors or camera movement onto the output. sound: Provides audio style or sound samples for synchronized generation.
This approach allows creative workflows familiar to CapCut users or that echo ByteDance’s robust editing platforms to migrate into an API-first generation environment effortlessly.
Example of references section
"references": [ "role": "style", "url": "https://cdn.example.com/paintstyle.jpg" , "role": "motion", "url": "https://cdn.example.com/dance-move.mp4" , "role": "sound", "url": "https://cdn.example.com/background-track.mp3" ]Native Synchronized Audio in the Same Generation Pass
Unlike legacy systems that generate video and audio separately (often requiring tedious syncing), Seedance 2.0 integrates audio generation into the very same rendering pass. By toggling generate_audio: true inside the audio object of seedanceParams, the API outputs a seamlessly synchronized audiovisual file.
This means you get natural sound morphing and direct correlation with visual events — perfect for add-tech applications requiring spot-on audiovisual alignment.
Director-Style Camera Movement via Prompt Language
Seedance 2.0 shines with its natural-language-driven camera direction. Instead of complicated keyframe definitions, you simply describe desired camera moves in the prompt:
- “Slow tilt up to reveal the horizon” “Dynamic zoom-in with a slight left pan” “Smooth dolly forward through crowd”
These commands empower creators to achieve complex cinematography with just text, a significant leap compared to many existing video APIs. When combined with the motion reference role, you can mix precise real-world motion with creative cinematic effects.
Job Tracking and Retrieval
Once you submit your generation request:
POST https://api.apiframe.ai/v2/videos/generateYou receive a jobId to track the asynchronous process. Polling or webhook callbacks connect with:
GET https://api.apiframe.ai/v2/jobs/jobIdThis endpoint returns status codes and result URLs:
Status Code Status Meaning Notes 200 Success Payload includes asset URLs and metadata. 202 In Progress Generation underway. 400 Bad Request Malformed parameters or missing required fields. 500 Server Error Internal failure, retry suggested.Billing Model: Priced Per Second of Video Output
An important consideration for production usage is Apiframe’s billing, which charges per second of generated video output. This transparent model encourages users to optimize their video lengths and resolutions, aligning cost directly with deliverable content Get more info rather than compute time or input complexity.
For instance, a 10-second video at 1080p with native audio will bill the equivalent of 10 seconds regardless of how many multimodal references or stylistic layers you applied. This predictability is a plus, especially compared to platforms that bill by total GPU compute time.
Summary - Where Are the Seedance 2.0 Fields?
To recap:
- The seedanceParams schema lives inside the POST https://api.apiframe.ai/v2/videos/generate request body. Fields like prompt, references (with explicit roles), duration, and audio options control every aspect of generation. Use GET https://api.apiframe.ai/v2/jobs/id to retrieve video assets and status. Billing is straightforward — per second of final video output. Camera movement is controlled natively via prompt language, integrated with motion references. This API unifies text, image, and video references, simplifying complex multimedia workflows for creators and ad-tech teams alike.
Closing Thoughts
The Seedance 2.0 API represents a significant step forward for asynchronous, API-first creative tooling, akin to what you find at ByteDance and CapCut but with web-native developer experience baked in. By carefully analyzing the seedanceParams input schema within the Apiframe docs, developers can unlock powerful narrative-driven and stylistic video generation with simple yet expressive inputs.
If you’re considering integrating AI video generation into your workflow, whether for advertising, entertainment, or social content platforms, Seedance 2.0 offers an API that scales easily and bills transparently, making creative experimentation efficient and predictable.