Skip to main content
POST
/
v1
/
videos
/
generations
curl --request POST \
  --url https://toapis.com/v1/videos/generations \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "doubao-seedance-2-0",
    "prompt": "Use video 1 for the POV composition throughout, use audio 1 as the background bed, start from image 1, and end on image 2 with a fresh commercial tone.",
    "duration": 11,
    "aspect_ratio": "16:9",
    "image_with_roles": [
      {"url": "https://example.com/ref-image-1.jpg", "role": "reference_image"},
      {"url": "https://example.com/ref-image-2.jpg", "role": "reference_image"}
    ],
    "video_with_roles": [
      {"url": "https://example.com/ref-video-1.mp4", "role": "reference_video"}
    ],
    "audio_with_roles": [
      {"url": "https://example.com/ref-audio-1.mp3", "role": "reference_audio"}
    ],
    "metadata": {
      "resolution": "720p",
      "generate_audio": true
    }
  }'

{
  "id": "<string>",
  "object": "<string>",
  "model": "<string>",
  "status": "<string>",
  "progress": 123,
  "created_at": 123,
  "metadata": {}
}
  • ByteDance’s next-generation video generation models
  • Supports doubao-seedance-2-0 and doubao-seedance-2-0-fast
  • Supports text-to-video, first-frame image-to-video, first-and-last-frame image-to-video, and multimodal reference-to-video
  • Supports reference images, reference videos, and reference audio
  • Supports synced audio generation, web search tools, and returning the last frame
  • Async task workflow with task ID based status queries

​
Authorizations

​
Authorization
string
required
All endpoints require Bearer Token authenticationGet your API Key from the API Key Management PageAdd to the request header:
Authorization: Bearer YOUR_API_KEY

​
Body

​
model
string
default:"doubao-seedance-2-0"
required
Video generation model nameAvailable models:
  • doubao-seedance-2-0 - Standard version focused on higher quality, supports 4-15 second output
  • doubao-seedance-2-0-fast - Faster version for preview and iteration, supports 4-12 second output
​
prompt
string
Video promptSupports both Chinese and English input. Describe scene, camera movement, subject actions, style, and audio atmosphere as clearly as possible.Recommendations:
  • Keep Chinese prompts within about 500 characters
  • Keep English prompts within about 1000 words
  • When referring to uploaded assets, use labels like “image 1”, “video 1”, or “audio 1”
Example: "Use video 1 for the POV composition throughout, start from image 1, end on image 2, and preserve the rhythm and mood from audio 1"
​
duration
integer
default:5
Video duration in secondsAllowed values:
  • doubao-seedance-2-0: 4-15
  • doubao-seedance-2-0-fast: 4-12
  • -1: automatic duration selected by the model
doubao-seedance-2-0-fast does not support durations longer than 12 seconds.
​
aspect_ratio
string
default:"adaptive"
Video aspect ratioOptions:
  • 21:9
  • 16:9
  • 4:3
  • 1:1
  • 3:4
  • 9:16
  • adaptive
adaptive behavior:
  • Text-to-video: the model chooses the most suitable ratio from the prompt
  • First-frame or first-and-last-frame image-to-video: derived from the first frame
  • Multimodal reference-to-video: typically prioritizes reference video, then reference image
​
image_urls
string[]
Image URL array in compatibility modeimage_with_roles is recommended for explicit control.image_urls and image_with_roles should not be used together.
​
image_with_roles
array
Image array with rolesSupported patterns:
  • First-frame image-to-video: one first_frame
  • First-and-last-frame image-to-video: one first_frame plus one last_frame
  • Multimodal reference-to-video: reference_image entries, 1-9 items
Image requirements:
  • Formats: jpeg, png, webp, bmp, tiff, gif
  • Per-image size: less than 30MB
  • Total request size: recommended within 64MB
  • Aspect ratio: about 0.4 to 2.5
  • Dimensions: about 300px to 6000px
  • First-frame and first-and-last-frame modes cannot be mixed with reference_image, reference_video, or reference_audio
  • Only one first_frame and one last_frame are allowed
  • In multimodal reference mode, all images should use reference_image
​
video_with_roles
array
Video array with rolesCurrently only reference_video is supported for multimodal reference mode.Video requirements:
  • Formats: mp4, mov
  • Resolution: 480p or 720p
  • Per-video duration: 2-15 seconds
  • Maximum count: 3
  • Total reference video duration: no more than 15 seconds
  • Per-video size: less than 50MB
  • Frame rate: about 24-60 FPS
​
audio_with_roles
array
Audio array with rolesCurrently only reference_audio is supported for multimodal reference mode.Audio requirements:
  • Formats: wav, mp3
  • Per-audio duration: 2-15 seconds
  • Maximum count: 3
  • Total reference audio duration: no more than 15 seconds
  • Per-audio size: less than 15MB
audio_with_roles cannot be used alone. At least one image or video reference is also required.
​
metadata
object
Extended parameters

​
Input Combination Rules

Typical supported combinations:
  • Text only: text-to-video
  • Text + one first-frame image: first-frame image-to-video
  • Text + first-frame image + last-frame image: first-and-last-frame image-to-video
  • Text + reference images: multimodal reference-to-video
  • Text + reference videos: video-guided reference-to-video
  • Text + reference images + reference audio: multimodal reference-to-video
  • Text + reference images + reference videos + reference audio: multimodal reference-to-video
These three modes are mutually exclusive:
  • First-frame image-to-video
  • First-and-last-frame image-to-video
  • Multimodal reference-to-video
If you need strict first-frame and last-frame control, prefer first_frame and last_frame. If you need broader reference guidance, use reference_image, reference_video, and reference_audio.

​
Resolution and Aspect Ratio Pixel Map

ResolutionAspect RatioPixel Size
480p16:9864x496
480p4:3752x560
480p1:1640x640
480p3:4560x752
480p9:16496x864
480p21:9992x432
720p16:91280x720
720p4:31112x834
720p1:1960x960
720p3:4834x1112
720p9:16720x1280
720p21:91470x630

​
Capabilities and Constraints

ItemSeedance 2.0Seedance 2.0 Fast
PositioningHigher qualityFaster generation and lower cost
Duration4-15 seconds, or -1 auto4-12 seconds, or -1 auto
Resolution480p / 720p480p / 720p
Image rolesfirst_frame / last_frame / reference_imagefirst_frame / last_frame / reference_image
Video rolesreference_videoreference_video
Audio rolesreference_audioreference_audio
Audio generationmetadata.generate_audiometadata.generate_audio
Toolsmetadata.toolsmetadata.tools
Return last framemetadata.return_last_framemetadata.return_last_frame
Pricing is billed per second. Final display prices may vary by model version, resolution, and marketplace strategy. Please refer to the model pricing page.

​
Response

​
id
string
Unique task identifier for status queries
​
object
string
Object type, always generation.task
​
model
string
Model name used
​
status
string
Task status
  • queued - Queued
  • in_progress - Processing
  • completed - Completed successfully
  • failed - Failed
​
progress
integer
Task progress percentage (0-100)
​
created_at
integer
Task creation timestamp (Unix timestamp)
​
metadata
object
Task metadata
Video generation is asynchronous. The create call returns a task ID, and you can use Get Video Task Status to poll for progress and results.
curl --request POST \
  --url https://toapis.com/v1/videos/generations \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "doubao-seedance-2-0",
    "prompt": "Use video 1 for the POV composition throughout, use audio 1 as the background bed, start from image 1, and end on image 2 with a fresh commercial tone.",
    "duration": 11,
    "aspect_ratio": "16:9",
    "image_with_roles": [
      {"url": "https://example.com/ref-image-1.jpg", "role": "reference_image"},
      {"url": "https://example.com/ref-image-2.jpg", "role": "reference_image"}
    ],
    "video_with_roles": [
      {"url": "https://example.com/ref-video-1.mp4", "role": "reference_video"}
    ],
    "audio_with_roles": [
      {"url": "https://example.com/ref-audio-1.mp3", "role": "reference_audio"}
    ],
    "metadata": {
      "resolution": "720p",
      "generate_audio": true
    }
  }'