curl --request POST \
  --url https://api.goenhance.ai/api/v1/videos/generations \
  --header 'Authorization: <authorization>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "seedance-2-0",
  "prompt": "A cat plays piano in moonlight, cinematic close-up",
  "duration": 8,
  "quality": "720p",
  "aspect_ratio": "16:9",
  "generate_audio": true
}
'

{
  "code": 0,
  "msg": "Success",
  "data": {
    "img_uuid": "c12b656c-747a-44fd-9c80-add79b0c52d5"
  }
}

Seedance

Seedance 2.5 Reference to Video

Seedance 2.5 reference-to-video: drive generation with reference images, videos and audio. Reference the assets inside the prompt as @image1, @video1, @audio1 — the array order is what those numbers refer to.

Pricing — per second of generated video, in tokens (USD at $0.02/token):

resolution	tokens/s	USD/s
480p	8.515	$0.1703
720p	18.265	$0.3653

Same price as Seedance 2.5 standard; all three modes cost the same.

Differences from Seedance 2.0 reference-to-video:

Far higher asset limits: 30 images, 10 videos, 10 audio clips (2.0 allows 9/3/3).
Reference video and audio may total 30 seconds each (2.0 allows 15).
Audio alone is enough — 2.0 required at least one image or video.
Only 480p and 720p; duration up to 30 seconds.

Returns an img_uuid; poll GET /api/v1/jobs/detail (or use custom_callback_url) to get the generated video.

⚠️ When ref_videos is used, the input video counts toward the bill. The charge becomes (total input video seconds + output duration) × rate × 0.8 — the reference footage is billed at the same per-second rate as the output, with a 20% discount applied to the whole thing. Input length is measured from your files before any tokens are deducted, so a 6s reference video on a 5s generation is billed as 11 seconds × 0.8, not 5.

Requests with only ref_imgs (no ref_videos) are unaffected: output duration × rate.

POST

api

videos

generations

curl --request POST \
  --url https://api.goenhance.ai/api/v1/videos/generations \
  --header 'Authorization: <authorization>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "seedance-2-0",
  "prompt": "A cat plays piano in moonlight, cinematic close-up",
  "duration": 8,
  "quality": "720p",
  "aspect_ratio": "16:9",
  "generate_audio": true
}
'

{
  "code": 0,
  "msg": "Success",
  "data": {
    "img_uuid": "c12b656c-747a-44fd-9c80-add79b0c52d5"
  }
}

Headers

Authorization

string

required

Body

application/json

model

enum<string>

required

Model name. Must be seedance-2-5-refs.

Available options:

seedance-2-5-refs

prompt

string

required

Text prompt. Chinese: recommended up to 500 characters; English: recommended up to 1000 words.

ref_imgs

string<uri>[]

Reference images (up to 30). jpeg/png/webp, ≤30MB each, 300–6000px, aspect ratio 0.4–2.5. Referenced in the prompt as @image1, @image2, … in array order.

Maximum array length: 30

ref_videos

string<uri>[]

Reference videos (up to 10). mp4/mov, 2–30s each and ≤30s in total, 24–60 fps, ≤200MB each. Longer clips are trimmed automatically — the 30s budget is split evenly across them (1 clip keeps 30s, 3 clips get 10s each). Referenced as @video1, @video2, … in array order. ⚠️ Using this field changes how the request is priced — see the pricing note above.

Maximum array length: 10

ref_audios

string<uri>[]

Reference audio (up to 10). wav/mp3, 2–30s each and ≤30s in total, ≤15MB each. Not transcoded — clips exceeding the total budget are rejected upstream. Referenced as @audio1, @audio2, … in array order. At least one of ref_imgs / ref_videos / ref_audios is required; unlike Seedance 2.0, audio on its own is accepted.

Maximum array length: 10

duration

integer

default:5

Any integer from 4 to 30 seconds.

Required range: 4 <= x <= 30

quality

enum<string>

default:720p

Output resolution. resolution is accepted as a compatible alias.

Available options:

480p,

720p

aspect_ratio

enum<string>

default:16:9

Output aspect ratio. Reference-to-video accepts every value below. ratio is accepted as a compatible alias.

Available options:

16:9,

9:16,

1:1,

4:3,

3:4,

21:9,

adaptive

generate_audio

boolean

default:true

Generate synchronized audio at no additional charge.

callback_url

string<uri>

Optional public HTTPS callback URL. custom_callback_url is accepted as a compatible alias.

Response

Task submitted or application-level error.

code

integer

required

msg

string

required

data

object

Show child attributes

Getting started

API

Seedance 2.5 Reference to Video

Headers

Body

Response