Feature · AI video

AI video generation API with Seedance

Generate short videos from text through an asynchronous API. Submit a prompt, choose duration, resolution and ratio, then poll the task until the video URL is ready.

Short answer. Submit POST /v1/videos with a prompt and optional duration (2–12s), resolution (480p/720p/1080p) and ratio (16:9, 9:16 or 1:1). The default model is plugsky-video → Seedance 1.0 Pro Fast. The response contains a task id and a polling_url; poll GET /v1/videos/{id} until video_url is populated. Video is paid-only.

Endpoint and request shape

FieldValuesNotes
promptTextDescribe subject, motion, camera and mood
modelplugsky-videoDefault maps to Seedance 1.0 Pro Fast
duration2–12 secondsShorter clips are cheaper and faster
resolution480p, 720p, 1080pMatch the destination platform
ratio16:9, 9:16, 1:1Landscape, vertical or square

Submit a video job

curl https://api.plugsky.com/v1/videos \
  -H "Authorization: Bearer $PLUGSKY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A slow push-in on a ceramic coffee cup as steam rises, warm morning light, shallow depth of field",
    "model": "plugsky-video",
    "duration": 6,
    "resolution": "720p",
    "ratio": "16:9"
  }'

The response returns a task id and a polling_url. Store both. Do not hold the original HTTP request open while the video renders.

Poll the result

curl https://api.plugsky.com/v1/videos/TASK_ID \
  -H "Authorization: Bearer $PLUGSKY_API_KEY"

The status moves through queued, running, succeeded or failed. On success, the response contains video_url. On failure, record the error and do not retry blindly; a malformed prompt or an unsupported ratio should fail fast.

A simple polling worker

import time
import requests

API = "https://api.plugsky.com/v1"
HEADERS = {"Authorization": "Bearer YOUR_PLUGSKY_KEY"}

job = requests.post(f"{API}/videos", headers=HEADERS, json={
    "prompt": "A product slowly rotating on a seamless background, studio light",
    "duration": 6,
    "resolution": "720p",
    "ratio": "1:1",
}).json()

for _ in range(60):
    status = requests.get(f"{API}/videos/{job['id']}", headers=HEADERS).json()
    if status["status"] == "succeeded":
        print(status["video_url"])
        break
    if status["status"] == "failed":
        raise RuntimeError(status)
    time.sleep(5)
else:
    raise TimeoutError(job["id"])

Video seconds by plan

PlanFair-use video seconds / monthBest for
Hobby $860Ten 6-second clips or a small test batch
Starter $20120Twenty 6-second clips or A/B variants
Builder $60300Fifty 6-second clips or a weekly content calendar
Scale $120600One hundred 6-second clips or high-volume ads

The free plan is excluded from video generation. Plan video seconds are a fair-use allowance, not a promise of unlimited rendering; longer clips and higher resolutions consume more capacity. See pricing for current terms.

Use cases that fit short clips

  • Social ads: 6–10 second vertical product shots with a clear motion beat.
  • Product demos: a slow rotation or push-in that shows texture and scale.
  • UGC-style variants: multiple backgrounds or camera moves for the same product.
  • Explainers: a short visual metaphor to open an article or landing page.

Prompt patterns

  • Subject + motion: “a ceramic cup, steam rising, slow push-in”.
  • Camera: push-in, orbit, handheld, static tripod.
  • Light: warm morning window light, hard studio strobe, overcast.
  • Duration: plan the shot for 4–8 seconds; longer clips need a reason.
  • Avoid text: generated on-screen text is unreliable; add typography in the editor.

Limitations

  • Clips are 2–12 seconds; this is not a long-form video editor.
  • Generation is asynchronous and can queue; design the UX around a job status.
  • Review outputs for brand, copyright, likeness and factual claims before publishing.
  • Audio is not part of the request contract; add voiceover or music in post.

FAQ

Which model powers video generation?

plugsky-video, mapping to Seedance 1.0 Pro Fast. The request returns a task id and polling URL.

How long can a generated video be?

2–12 seconds, at 480p, 720p or 1080p, in 16:9, 9:16 or 1:1.

How do I get the finished video?

Poll GET /v1/videos/{id} until the status is succeeded; the response contains video_url.

How many video seconds are included?

Hobby 60, Starter 120, Builder 300 and Scale 600 fair-use seconds per month. The free plan is excluded.

Is video generation synchronous?

No. Submit the job, store the id, and poll or use a worker until it finishes.

Reviewed 20 November 2026. Endpoint behaviour and video-second allowances were checked against the Plugsky API documentation on that date. Review generated video for brand, copyright and likeness before publication.