Seedance 2.5
Seedance 2.5 makes cinematic clips of 4 to 30 seconds with synchronized audio from text, a first frame, or a mix of image, video and audio references. It is best at multi-beat storytelling, dialogue with lip-sync, and product motion in one long take.
video · model seedance-2.5 · catalog model-catalog@2026-10-06.v5 · pricing customer-pricing@2026-11-05.v15.Endpoint
POST https://api.roasos.com/v1/video/seedance-2.5Send your API key as a bearer token. The key needs the scope api:write. Every paid request also needs an Idempotency-Key header.
The envelope fields quote_only, quote_id, webhook_url and wait_timeout_seconds behave the same on each model endpoint. Refer to Jobs and Quotes and credits.
Good at
Seedance 2.5 is the premium video model of the roster. It makes one continuous take of up to 30 seconds with sound effects, ambience and lip-synced speech; the audio has no surcharge. It follows beat-by-beat prompts with timestamps, and it keeps a product, a set and a palette stable across cuts. Reference mode binds images, videos and audio files to ordinal tags, so a caller can lock a product look, copy a motion, or use a voice track. It accepts a human face in a reference image, a real photo or a saved avatar. roasOS runs this model on a provider route that accepts faces.
Pricing
The unit is one clip. Pricing key: seedance-2.5.
| Request | Credits |
|---|---|
4s_720p | 12 |
12s_720p | 36 |
30s_720p | 90 |
4s_720p_no_audio | 12 |
The credits come from the roasOS pricing table. No other field changes the price. To get the exact price of your request, send it with quote_only: true first.
Modes
text-to-videoimage-to-videoreference-to-video
roasOS selects the mode from the media fields of your request. You do not send a mode.
Provider endpoints: bytedance/seedance-2.5/image-to-video · bytedance/seedance-2.5/reference-to-video · bytedance/seedance-2.5/text-to-video.
Request fields
| Field | Type | Required | Default | Rule |
|---|---|---|---|---|
aspect_ratio | enum | no | 16:9 | auto, 9:16, 1:1, 3:4, 4:3, 16:9, 21:9; only in text-to-video, reference-to-video |
audio_urls | string[] | no | — | ≤ 10 items; only in reference-to-video |
avatar_id | string | no | — | UUID; only in reference-to-video |
avatar_outfit | string | no | — | ≤ 500 characters |
duration_seconds | enum | no | 5 | auto, 4, 5, 6, 8, 10, 12, 15, 20, 25, 30; required in reference-to-video |
end_image_url | string | no | — | https URL; ≤ 4,096 characters; only in image-to-video |
extend_of_generation_id | string | no | — | UUID |
generate_audio | boolean | no | true | — |
image_url | string | no | — | https URL; ≤ 4,096 characters; only in image-to-video |
mode | enum | no | — | text-to-video, image-to-video, reference-to-video |
prompt | string | yes | — | ≥ 1 character; ≤ 10,000 characters |
quote_id | string | no | — | UUID |
quote_only | boolean | no | false | — |
references | string[] | no | — | ≤ 30 items; only in reference-to-video |
resolution | enum | no | 720p | 480p, 720p, 1080p |
scene_id | string | no | — | UUID |
video_urls | string[] | no | — | ≤ 10 items; only in reference-to-video |
wait_timeout_seconds | integer | no | 0 | 0–60 |
webhook_url | string | no | — | https URL; ≤ 4,096 characters |
The schema is strict. A field that is not in this table gets 400 unsupported_option. The field names are the roasOS role names. One role has one name in every family.
Media
This endpoint accepts these media roles:
avatar_id— identity; the saved character reference sheet, with a derived portrait only where the model requires a single face or source framereferences— style and appearance referencesimage_url— first frame (or the single source image)end_image_url— last framescene_id— first frame — a scene this workspace ownsvideo_urls— reference videosaudio_urls— reference or driving audio
| Role | Items | Maximum size | Content types | Duration |
|---|---|---|---|---|
audio_urls | ≤ 10 | 14.3 MB | audio/mpeg, audio/wav | 1.8–30.2 s |
end_image_url | ≤ 1 | 10 MB | image/jpeg, image/png, image/webp | — |
image_url | ≤ 1 | 10 MB | image/jpeg, image/png, image/webp | — |
references | ≤ 30 | 10 MB | image/jpeg, image/png, image/webp | — |
video_urls | ≤ 10 | 100 MB | video/mp4 | 2–30 s |
Every other role is refused: elements, product_id, product_images, reference_images, scene_image, video_url, audio_url, voice_id. A refused role gets 400 unsupported_option.
https://cdn.roasos.com that your workspace owns. Any other URL gets 400 reference_media_not_workspace_owned. Refer to Reference media and uploads to learn how to put a file there.Prompt
Limit 10,000 characters including the roasOS route-facts block (ROUTE-VERIFIED FEATURE FACTS: with one line for the first frame or the references). Tags: @Image1, @Video1, @Audio1, each once, in order.
- Order: Subject → Action → Camera → Style/lighting → Constraints. One primary movement per beat.
- Give each beat a start time:
[00:00] … [00:05] …. - Write dialogue as
She says: "…"with the delivery; name the mic source ("direct phone mic, room ambience, no music"). - Dialogue pacing: 1–8 words ≈ 4–5 s, 9–15 ≈ 6–8 s, 16–25 ≈ 9–12 s, 26–35 ≈ 13–15 s.
- Avoid: cinematic, professional, stunning, 8k, studio, perfect. Say "documentary", "handheld", "phone-shot".
- Motion needs intensity and degree adverbs ("slowly picks up the bottle with her right hand and turns the label to camera").
- Lock a product: "The product from @Image2 must remain visually unchanged in every shot".
- A face in an image reference is accepted. An
avatar_idappends the avatar's front and full-body shots as the trailing@ImageNentries. - If the provider refuses a face, the run fails with
provider_rejects_face_references. roasOS refunds the reservation.
Not exposed
These catalog entries describe omitted inputs, pinned values and source constraints. They do not imply that the provider accepts every listed field.
duration_seconds— these values are not available:7,9,11,13,14,16,17,18,19,21,22,23,24,26,27,28,29.video_urls— these values are not available:MOV containers.references— these values are not available:BMP,TIFF,GIF,HEIC,HEIF.seed— pinned tooutput only.end_user_id— not accepted by this public endpoint.
Result
The result kind is video. Always present: url.
{
"data": {
"id": "gen_…",
"kind": "video",
"model": "seedance-2.5",
"status": "completed",
"result": {
"url": "https://cdn.roasos.com/workspaces/ws_01/result"
}
}
}The files are immutable objects on https://cdn.roasos.com. A result URL is a valid media input for your next request.
Example
The three steps are always the same: quote, run, poll.
# 1. Quote. Nothing is spent.
curl -X POST https://api.roasos.com/v1/video/seedance-2.5 \
-H "Authorization: Bearer roasos_sk_live_..." \
-H "Idempotency-Key: 0f1c9e2a-seedance-2.5-0001" \
-H "Content-Type: application/json" \
-d '{"quote_only":true,"prompt":"A woman holds the bottle at eye level in a bright kitchen, morning light, static camera."}'
# 200 { "data": { "kind": "quote", "quote_id": "q_7Hk...", "credits": 12, "expires_at": "...",
# "confirmation": { "model": "seedance-2.5", "example": "4s_720p" } } }
# 2. Run. Same body, same Idempotency-Key, quote_id in place of quote_only.
curl -X POST https://api.roasos.com/v1/video/seedance-2.5 \
-H "Authorization: Bearer roasos_sk_live_..." \
-H "Idempotency-Key: 0f1c9e2a-seedance-2.5-0001" \
-H "Content-Type: application/json" \
-d '{"quote_id":"q_7Hk...","prompt":"A woman holds the bottle at eye level in a bright kitchen, morning light, static camera."}'
# 201 { "data": { "id": "gen_...", "status": "queued", "credits_reserved": 12 } }
# 3. Poll, then read the result.
curl https://api.roasos.com/v1/jobs/gen_.../status -H "Authorization: Bearer roasos_sk_live_..."
curl https://api.roasos.com/v1/jobs/gen_.../result -H "Authorization: Bearer roasos_sk_live_..."The deprecated path
POST /v1/generations still reaches this model for one minor version. It is deprecated. This is how it maps:
POST /v1/generations { "type": "video", "model": "seedance-2.5", "options": { prompt, image, last_frame, reference_images, reference_videos, audio_urls, duration, resolution, aspect_ratio, generate_audio } }. Contract authority exact for all three modes. It returns the same job envelope. The alias stays one minor version after v1.1 ships.
See also
llms.txtfor this model — the same facts as one plain-text file for an agent.GET https://api.roasos.com/v1/catalog/seedance-2.5— the machine-readable form: enums, defaults, limits and prices. Refer to the API reference.- Jobs — how to poll a job and read its result.
- Provider endpoints:
bytedance/seedance-2.5/image-to-video·bytedance/seedance-2.5/reference-to-video·bytedance/seedance-2.5/text-to-video.
Seedance 2.0 Pro
Seedance 2.0 Pro makes cinematic clips of 4 to 15 seconds with synchronized audio at up to 4K, from text, a first frame, or image, video and audio references. It is best for the final cut of an ad and for hero product motion on a large screen.
Seedance 2.5 Edit
Seedance 2.5 Edit changes an existing clip by instruction. The output keeps the length, the framing and the timing of the source. Use it to put your product or your character into a clip that you want to keep.