Seedance 2.5 API Docs

Flagship cinema-grade text-to-video and first/last-frame image-to-video up to 30s and 1080p, with native synced audio included and 0.1s duration billing.

POST /v1/videos/generations

Public model IDs & modes

Flagship BytePlus cinema video generation supporting 4–30 second clips across 480p, 720p, and 1080p.

Pass a *-text-to-video ID for prompt generation, or an *-image-to-video ID with 1 start frame or 2 keyframes ([first_frame, last_frame]) in image_urls. Both modes share the same per-second rate, and native synchronized audio is included at no extra charge.

Public Model IDSupported ModeRequired Input
seedance-2.5-text-to-videoText-to-Video (4–30s)prompt
seedance-2.5-image-to-videoFirst/Last Frame Image-to-Video (4–30s)prompt + image_urls (1..2)

Request schema

Unknown top-level fields are rejected immediately with HTTP 400 invalid_parameter before any credits are reserved. On *-image-to-video models, omit aspect_ratio or set it to "adaptive".

FieldType & DefaultConstraints & Description
Authorization
RequiredHeader
string
Bearer cm_live_... server-side API key from your account.
Idempotency-Key
OptionalHeader
stringDefault: recommended
1 to 255 ASCII chars. Prevents duplicate submissions and charges for 24h.
model
Required
"seedance-2.5-text-to-video" | "seedance-2.5-image-to-video"
Select text-to-video or first/last-frame image-to-video public ID.
prompt
Required
string
Non-empty UTF-8 scene and motion prompt.
image_urls
Optional
string[] (1..2)Default: required on I2V
1 first frame or 2 first+last HTTPS frame URLs on image-to-video (0 extra fee).
quality
Optional
"480p" | "720p" | "1080p"Default: "720p"
Output video resolution tier; determines per-second credit rate.
duration
Optional
integer (4..30s)Default: 5
Requested clip duration (4..30s). Minimum 4.0s floor; settles on exact 0.1s actual output.
aspect_ratio
Optional
"16:9" | "9:16" | "1:1" | "4:3" | "3:4" | "21:9" | "adaptive"Default: "adaptive"
T2V accepts all 7 values; I2V accepts only "adaptive" (follows input frame).
generate_audio
Optional
booleanDefault: true
Native synchronized sound effects & dialogue — included at no extra charge.
output_format
Optional
"mp4" | "mov"Default: "mp4"
Output video container format returned via 30-day HTTPS URL.

Submit & poll examples

Video synthesis typically completes in 20–90 seconds. Poll GET /v1/tasks/:task_id every 3–5 seconds until status reaches completed or failed.

curl -sS -X POST "https://api.callmodels.dev/v1/videos/generations" \
  -H "Authorization: Bearer $CALLMODELS_API_KEY" \
  -H "Idempotency-Key: vid_t2v_20261115_01" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-text-to-video",
  "prompt": "A smooth tracking shot along wet volcanic coastline at sunrise with natural surf audio",
  "quality": "720p",
  "duration": 5,
  "aspect_ratio": "16:9",
  "generate_audio": true,
  "output_format": "mp4"
}'

Task response payloads

Completed video tasks return a 30-day signed MP4 URL on https://files.callmodels.dev together with the final cost_credits.

{
  "id": "task_01e561ba-859e-47fe-be79-60bc6f1d6499",
  "model": "seedance-2.5-text-to-video",
  "object": "video.generation.task",
  "status": "pending",
  "created_at": "2026-11-15T09:49:18.000Z"
}

Per-second pricing & 0.1s settlement

Submission temporarily holds duration × rate_per_second. Upon completion, billing settles to actual output duration rounded up to 0.1s (4.0s minimum, capped at requested duration) and returns any unused hold.

SpecificationCallmodels CreditsUSD Equiv.Official Ref.
480p · T2V or I2V · Synced audio included (5s = 70.31 Credits)14.061121 Credits / s$0.1406 / s—
720p · T2V or I2V · Synced audio included (5s = 151.19 Credits)30.238971 Credits / s$0.3024 / s—
1080p · T2V or I2V · Synced audio included (5s = 377.99 Credits)75.597426 Credits / s$0.7560 / s—