Bytedance Seedance 2.5 Reference To Video Turbo API Documentation

Bytedance Seedance 2.5 Reference To Video Turbo API Documentation

Playground

Seedance 2.5 Turbo Reference to Video creates cinematic videos with native audio from your own reference images, videos, and audio, keeping characters, products, motion, and sound consistent while a prompt directs the new shot. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

Seedance 2.5 Turbo Reference-to-Video guides generation with your own reference images, videos, and audio. Keep characters, products, scenes, motion, and soundtrack consistent while a prompt directs the new shot, with native synchronized audio and director-style control over camera, lighting, and pacing. The same reference inputs are also accepted by Seedance 2.5 Turbo Text-to-Video.

Why Choose This?

• Reference-guided generation
  Use reference images, videos, and audio to steer characters, products, composition, motion, and soundtrack.

• Consistent subjects
  Keep the same character, product, or scene across shots by attaching it as a reference.

• Native audio generation
  Generate synchronized audio together with the video in a single workflow.

• Director-style prompt control
  Control camera movement, lighting, performance, atmosphere, and visual style through prompts.

• Flexible output settings
  Choose aspect ratio, duration, resolution, and audio generation settings for different creative workflows.

Parameters

How to Use

1. Write your prompt — Describe the scene, subject, action, camera movement, lighting, mood, and visual style.
2. Add references optional — Use reference images for style, characters, or composition; reference videos for motion or pacing; and reference audio for soundtrack direction.
3. Choose aspect ratio — Select the output layout that matches your target format.
4. Choose resolution — Use 720p for lower-cost generation or 1080p for higher-resolution output.
5. Set duration — Choose a video length from 4 to 30 seconds.
6. Configure audio optional — Keep generate_audio enabled when synchronized audio is needed.
7. Submit — Generate the final video.

Pricing

Pricing is calculated upfront using your selected duration, plus reference-video charges when applicable.

Without reference videos

Output-only pricing:

With reference videos

Pricing has two parts:

• Normalized reference-video duration + requested generation duration: $0.11 per second.
• Output upscale: $0.02 per requested second at 720p or $0.04 per requested second at 1080p.

Each reference video is normalized to at least 2 seconds, each clip and the combined input are capped at 30 seconds, and the normalized combined input duration is rounded up to a whole second for billing.

Best Use Cases

• Turbo HD video generation — Generate 720p or 1080p videos from text prompts with faster turnaround.
• High-volume content production — Create cost-efficient high-resolution videos for large creative pipelines.
• Social media content — Produce vertical, square, landscape, or ultra-wide videos for different platforms.
• Ad creatives — Generate cinematic commercial concepts, campaign videos, and product scenes.
• Reference-guided generation — Use images, videos, or audio to guide style, motion, characters, or soundtrack direction.
• Creative prototyping — Quickly test scene ideas, camera movement, lighting, and mood before final production.

Pro Tips

• Write prompts like a film director, including subject, action, camera movement, lighting, mood, and atmosphere.
• Use 720p for lower-cost testing and 1080p for higher-resolution output.
• Start with shorter durations for quick iteration, then increase duration once the visual direction is confirmed.
• Use reference images when character, product, or style consistency matters.
• Use reference videos when motion, pacing, or camera movement matters.
• Use reference audio when soundtrack, ambience, or voice style matters.
• Use asset tags such as @Image 1, @Image 2, and @Audio 1 to reference uploaded assets in the prompt.
• Keep reference instructions clear so the model knows which asset controls style, character, motion, or audio direction.

Related Models

• Seedance 2.5 Reference-to-Video — Higher-quality text-to-video generation.
• Seedance 2.5 Image-to-Video Turbo — Generate videos from reference images with the turbo workflow.
• Seedance 2.0 Fast Reference-to-Video Turbo — Speed-optimized text-to-video generation.

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",
  "aspect_ratio": "16:9",
  "resolution": "720p",
  "duration": 5,
  "generate_audio": true
}
JSON
)

# 1. Submit the prediction once. Never retry this request automatically:
# a retry after an unclear failure can create a second, billed task.
CHECK='The task may already exist. Check History or the predictions list before submitting again.'
if ! SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  --connect-timeout 10 --max-time 60 \
  -X POST "https://api.wavespeed.ai/api/v3/bytedance/seedance-2.5/reference-to-video-turbo" \
  -H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
  -H "Content-Type: application/json" \
  -d "${REQUEST_BODY}"); then
  printf 'Submit failed. %s\n' "${CHECK}" >&2
  exit 1
fi

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 "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id. %s\n' "${CHECK}" >&2
  exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"

# 2. Poll until the prediction reaches a final status.
BODY_FILE=$(mktemp)
trap 'rm -f "${BODY_FILE}"' EXIT
while true; do
  # Transient errors (429, 5xx) are retried, honoring Retry-After.
  HTTP_CODE=$(curl --silent --show-error --fail-with-body \
    --connect-timeout 10 --max-time 60 --retry 5 --retry-delay 2 \
    --output "${BODY_FILE}" --write-out '%{http_code}' \
    "${RESULT_URL}" \
    -H "Authorization: Bearer ${WAVESPEED_API_KEY}" || true)
  case "${HTTP_CODE}" in
    2??) ;;
    000|429|5??)  # dropped connection or temporary error: keep polling
      printf 'Task %s: HTTP %s, polling again in 10s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      sleep 10; continue ;;
    *) printf 'Task %s: HTTP %s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
      cat "${BODY_FILE}" >&2; exit 1 ;;
  esac
  RESULT=$(jq 'if type == "object" and has("data") then .data else . end' "${BODY_FILE}")
  STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')

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

Parameters

Task Submission Parameters

Request Parameters

Every model also accepts the global request options enable_sync_mode and enable_base64_output; see their pages for details.

Response Parameters

Result Request Parameters

Result Response Parameters