Seedance (Kie.ai) icon

Seedance (Kie.ai)

Generate videos using Seedance via Kie.ai API

Actions3

Text-to-Video

AI-generated

Summary

Generate a video from a text prompt using the Seedance models via Kie.ai API.

Inputs

  • Model — Select the Seedance model for text-to-video generation from options like Seedance 2.0, Seedance 2.0 Fast, and others.
  • Prompt (required) — Enter the text prompt describing the video content to generate.
  • Reference Image URLs — Optionally provide up to 9 reference image URLs to guide the video generation (Seedance 2.0 models only).
  • Reference Video URLs — Optionally provide up to 3 reference video URLs to influence the output (Seedance 2.0 models only).
  • Reference Audio URLs — Optionally provide up to 3 reference audio clip URLs to assist generation (Seedance 2.0 models only).
  • Duration — Choose the video length from 4, 8, 12, or 15 seconds.
  • Aspect Ratio — Set the video aspect ratio from options including 16:9, 9:16, 1:1, and Adaptive.
  • Resolution — Select the video resolution between 480p and 720p (Seedance 2.0 models only).
  • Generate Audio — Enable or disable audio generation in the output video (Seedance 2.0 models only).
  • Return Last Frame — Choose whether to return the last video frame as a separate image (Seedance 2.0 models only).
  • Web Search — Enable web search augmentation during generation (Seedance 2.0 models only).
  • Seed — Set a numeric seed for deterministic output; 0 means random seed.
  • Reply URL — Provide a webhook URL to receive a callback upon task completion.
  • Reply Ref — Custom reference string passed back in webhook callback.
  • Negative Prompt — Specify elements or features to avoid in the generated video.
  • Captcha Token — Include reCAPTCHA token if required by the API.
  • Wait for Completion — Choose whether to wait for the generation task to complete before continuing; polling every 3 seconds up to 5 minutes.

Output shape

a single JSON object representing the creation task response or, if waiting for completion, the final task result including URLs to the generated video and related media

If 'Wait for Completion' is enabled, the node polls asynchronously until the job finishes, returning the final data including video URLs. If disabled, it returns the initial task creation response with a task ID which can be used later to query status.

Examples

Example 1: Create an 8-second 16:9 video with audio from a text prompt using the Seedance 2.0 model

Model=bytedance/seedance-2, Prompt='A tranquil forest at dawn', Duration=8, Aspect Ratio=16:9, Resolution=720p, Generate Audio=true, Wait for Completion=true

Links

Discussion