Seedance 2.5 Extend API

Extend video forward or backward, keeping look and motion.

~234.41s

POST /v2/seedance-2.5-extend · submit + poll

 1# pip install "segmind>=1.1.0"
 2# export SEGMIND_API_KEY="YOUR_API_KEY"
 3from segmind import SegmindClient, InferenceFailed, InferenceTimeout
 4
 5# Async (v2) — recommended for long-running / video models.
 6# run() blocks up to 600s; submit_async + job.wait(timeout=...) sets a longer
 7# deadline and keeps the request_id so you can re-poll later.
 8client = SegmindClient()                      # reads SEGMIND_API_KEY
 9payload = {
10    "video": "https://segmind-resources.s3.amazonaws.com/input/seedance-2.5-edit-source-739a4cc6.mp4",
11    "direction": "forward",
12    "prompt": "The wind-up tin robot keeps marching to the end of the bench, reaches the row of colourful wooden building blocks and bumps the stack with its chest; the blocks topple and tumble across the workshop bench as the robot's legs keep ticking. Same stiff clockwork gait, same bench and window light, natural workshop ambience with the clatter of falling blocks.",
13    "duration": 6,
14    "resolution": "720p",
15    "generate_audio": True,
16}
17job = client.submit_async("seedance-2.5-extend", **payload)
18print(job.request_id)                         # available immediately
19try:
20    result = job.wait(timeout=900, interval=2.0)
21    print(result["status"])                  # COMPLETED
22    print(result.get("output"))              # model output (e.g. video URL)
23except InferenceTimeout as e:
24    print("still running:", e.request_id)    # re-poll later with this id
25except InferenceFailed as e:
26    print("failed:", e.detail)
27
28# Fast models (<=600s) can use the one-liner instead:
29# result = segmind.run("seedance-2.5-extend", **payload)

 1# pip install "segmind>=1.1.0"
 2# export SEGMIND_API_KEY="YOUR_API_KEY"
 3from segmind import SegmindClient, InferenceFailed, InferenceTimeout
 4
 5# Async (v2) — recommended for long-running / video models.
 6# run() blocks up to 600s; submit_async + job.wait(timeout=...) sets a longer
 7# deadline and keeps the request_id so you can re-poll later.
 8client = SegmindClient()                      # reads SEGMIND_API_KEY
 9payload = {
10    "video": "https://segmind-resources.s3.amazonaws.com/input/seedance-2.5-edit-source-739a4cc6.mp4",
11    "direction": "forward",
12    "prompt": "The wind-up tin robot keeps marching to the end of the bench, reaches the row of colourful wooden building blocks and bumps the stack with its chest; the blocks topple and tumble across the workshop bench as the robot's legs keep ticking. Same stiff clockwork gait, same bench and window light, natural workshop ambience with the clatter of falling blocks.",
13    "duration": 6,
14    "resolution": "720p",
15    "generate_audio": True,
16}
17job = client.submit_async("seedance-2.5-extend", **payload)
18print(job.request_id)                         # available immediately
19try:
20    result = job.wait(timeout=900, interval=2.0)
21    print(result["status"])                  # COMPLETED
22    print(result.get("output"))              # model output (e.g. video URL)
23except InferenceTimeout as e:
24    print("still running:", e.request_id)    # re-poll later with this id
25except InferenceFailed as e:
26    print("failed:", e.detail)
27
28# Fast models (<=600s) can use the one-liner instead:
29# result = segmind.run("seedance-2.5-extend", **payload)

API Endpoint

POSThttps://api.segmind.com/v1/seedance-2.5-extend

Parameters

directionrequired

string

Which way the clip grows: forward continues it after the last frame, backward adds what happens before the first. Use forward to continue a scene, backward for an establishing lead-in.

Default: "forward"

Allowed values :

Forward (continue the clip)→"forward"

Backward (what happened before)→"backward"

promptrequired

string

What happens in the new section of @Video 1; the direction is written in for you and the output is only the new footage. Describe one clear beat and cite extra references as @Image 1, @Video 2, @Audio 1.

videorequired

string (uri)

The clip to extend, cited as @Video 1: 2-30 s, mp4 or mov, 480p-4k, up to 200 MB. The output keeps its aspect ratio; real human faces in the source are blocked by the content policy.

auto_resizeoptional

boolean

Automatically resizes a reference image outside the accepted input range (300-6000px per side, ratio 0.4-2.5, 30 MB) so the request succeeds, preserving aspect ratio. Leave on; set false to have out-of-range images rejected instead.

Default: true

bitrate_modeoptional

string

Encoding bitrate: standard for smaller files, high for maximum fidelity. Use high for delivery masters, standard for quick previews.

Default: "standard"

Allowed values :

Standard→"standard"

High→"high"

durationoptional

integer

Length of the new footage in seconds (4-30), or -1 to let the model choose (about 6 s for a 5 s clip). Set a value for a precise beat; the source clip is billed on top.

Default: -1

Allowed values (28 total):

Auto (use for video edits)→-1

4s→4

5s→5

6s→6

7s→7

8s→8

9s→9

10s→10

11s→11

12s→12

+18 more

generate_audiooptional

boolean

true continues sound for the new footage in the clip's own style; false delivers no audio track. Leave on to carry the soundtrack across the join.

Default: true

output_codecoptional

string

Codec of the delivered file. 'auto' returns the model's own encode untouched - for seedance-2.5 at 1080p that is 10-bit HEVC, the highest-fidelity master, which Chrome and Firefox on Windows cannot decode and will play as audio over a blank frame. 'h264' re-encodes it to H.264 High 8-bit yuv420p, which plays everywhere, at the cost of one additional generation of compression and half the bit depth; pick it for browser embeds and previews, and leave it on 'auto' when the file goes into an edit or grade. Output that is already H.264 (720p, 480p, the 2.0 series) is never re-encoded either way. Adds a few seconds when it applies, and does not affect price.

Default: "auto"

Allowed values :

"auto""h264"

output_formatoptional

string

mp4 for best compatibility; mov (higher colour fidelity) for editing round-trips. Pick mov when the result feeds another edit or grade.

Default: "mp4"

Allowed values :

"mp4""mov"

reference_audiosoptional

string[]

Up to 10 MP3s (2-30 s each, 30 s combined), cited as @Audio 1. Use to drive voice, music or beat-matched pacing.

reference_imagesoptional

string[]

Up to 30 images for character or style consistency, cited as @Image 1. Supply reference art to keep a continued subject on-model.

reference_videosoptional

string[]

Up to 9 extra clips besides the source (2-30 s each, 30 s combined), cited as @Video 2 onward. Add for motion or style transfer; billed at the video-input rate.

resolutionoptional

string

Output resolution: 480p, 720p or 1080p at 24 fps. Use 480p for drafts, 720p for final renders, 1080p (10-bit) for highest fidelity.

Default: "720p"

Allowed values :

"480p""720p""1080p"

return_last_frameoptional

boolean

Returns the final frame as an image URL (video.last_frame_url). Enable to chain this extension into the next shot.

Default: false

seedoptional

integer

Reproducibility seed (-1 to 2147483647). Use -1 for a random result, fix a value to iterate the same continuation.

Default: 42Range: -1 - 2147483647

skip_moderationoptional

boolean

Bypasses the content pre-filter for false positives. Leave off unless artistic content is wrongly blocked.

Default: true

Response Type

Returns: Video

Asynchronous requests (v2)

Use Async for video, long-running (>~60s), or high-concurrency workloads; Sync is simplest for fast image & LLM calls. Async submits a request and you poll it to completion.

1
POST /v2/seedance-2.5-extend
Submit — returns request_id, status_url, response_url
2
GET /v2/requests/{id}/status
Poll — until COMPLETED or FAILED
3
GET /v2/requests/{id}
Result — final response body

Status states

QUEUED— Accepted, waiting for a worker

PROCESSING— Running on a worker

COMPLETED— Done — result body is ready

FAILED— Errored (incl. content/RAI blocks)

A FAILED request is served as HTTP 422 — the body still carries the error detail.
An unknown or expired request_id returns HTTP 404.
Results are retained for 1 hour, then expire.
Content / RAI blocks surface as FAILED, not a separate state.
Track completion by polling the status endpoint.

Common Error Codes

The API returns standard HTTP status codes. Detailed error messages are provided in the response body.

400

Bad Request

Invalid parameters or request format

401

Unauthorized

Missing or invalid API key

403

Forbidden

Insufficient permissions

404

Not Found

Model or endpoint not found

406

Insufficient Credits

Not enough credits to process request

429

Rate Limited

Too many requests

500

Server Error

Internal server error

502

Bad Gateway

Service temporarily unavailable

504

Timeout

Request timed out