Bytedance Seedance 2.0 Video Extend API Documentation

Bytedance Seedance 2.0 Video Extend API Documentation

Playground

Seedance 2.0 (Video-Extend) extends an input video with a new cinematic continuation generated from its last frame and a natural-language prompt. Ready-to-use REST API, best performance, no coldstarts, affordable pricing.

Features

Seedance 2.0 Video-Extend appends a new cinematic continuation to an existing video. The model generates a fresh segment from the input video’s last frame and a natural-language prompt, then the original and new segment are concatenated into a single output.

Key Features

• Seamless continuation — Generation starts from the input video’s last frame so the join is visually consistent.
• Optional target last frame — Provide last_image to steer the new segment toward a specific end frame.
• Director-level control — Camera movement, lighting, shadows, and character performance via prompts.
• Native audio synchronization — Audio from both the original video and the new segment are preserved in the final output.

Parameters

How to Use

1. Upload the input video. The new segment will be appended to it.
2. Write the prompt. Describe the continuation in cinematic detail.
3. (Optional) Upload a target last frame. The new segment will interpolate to this image.
4. Pick a duration and resolution.
5. Run. Receive the original video plus the new segment concatenated into one output.

Writing Effective Prompts

Extension here continues your video by generating a new segment from its final frame and stitching it on. So your prompt describes what happens next, starting from that last frame — the same way you’d direct an image-to-video shot. (This differs from native extension: don’t write “continue” or “extend backward” triggers; just describe the new action.)

Direct the new segment

Describe the action, motion, and camera work that begin from the final frame. Don’t re-describe what’s already on screen — build forward from it.

From the final frame, the surfer rides the wave to shore; the camera pulls back to a wide shot.

Keep it continuous

Match the motion, camera, and pacing of the source so the seam is invisible. For longer extensions, use a timed shot list.

> 0-3s: The bee lifts off the flower, the camera following in a slow macro pull-back. > 3-5s: It drifts to a second flower and lands; pollen catches the light as it settles.

Camera language

Write these directly; the model understands them:

• Shot size — extreme wide, wide, medium, medium close-up, close-up
• Movement — push in, pull out, pan, track, follow, orbit, tilt up, handheld shake
• Angle — low angle, overhead, eye-level, first-person
• Techniques — one-shot / long take, dolly zoom, bullet time, speed ramp

For a niche term, add a plain-language gloss: “rack focus: the sharp foreground softens as the background comes into focus.”

Negative control

Positive descriptions work best, but you can suppress subtitles and audio:

• “No subtitles.”
• “No BGM — ambient and action sounds only.”
• “No audio.”

Weak vs. strong

Pricing

Billed per second of the new segment only. Resolution multipliers match Seedance 2.0 image-to-video.

Examples (5s extension):

Notes

• Pricing is based on the new segment length only — the input video is not re-billed.
• Aspect ratio of the new segment matches the input video’s last frame automatically.
• Native audio generation is included for the new segment; the original video’s audio is preserved.

Related Models

• Seedance 2.0 Image-to-Video — Generate video from a reference image and prompt.
• Seedance 2.0 Video-Edit — Edit an existing video instead of extending it.
• Seedance 2.0 Fast Video-Extend — Faster, cheaper variant.

Authentication

For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result

set -euo pipefail

export WAVESPEED_API_KEY="your-api-key"

REQUEST_BODY=$(cat <<'JSON'
{
  "prompt": "A cinematic ocean wave at sunrise, highly detailed",
  "video": "https://interactive-examples.mdn.mozilla.net/media/cc0-videos/flower.mp4",
  "resolution": "720p",
  "duration": 5,
  "enable_web_search": false,
  "generate_audio": true
}
JSON
)

# 1. Submit the prediction.
SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  -X POST "https://api.wavespeed.ai/api/v3/bytedance/seedance-2.0/video-extend" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}")

TASK=$(printf '%s' "${SUBMIT_RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
PREDICTION_ID=$(printf '%s' "${TASK}" | jq -r '.id // empty')
if [ -z "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id
' >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 2. Poll until the prediction finishes.
while true; do
  RESPONSE=$(curl --silent --show-error --fail-with-body \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}")
  RESULT=$(printf '%s' "${RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

  case "${STATUS}" in
    completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
    failed|cancelled|timeout|deleted) printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    *) sleep 2 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Response Parameters

Result Request Parameters

Result Response Parameters