seedinteger
Random seed for reproducible results. Omit or use -1 for a random seed.
Minimum: -1
Maximum: 4294967295
bytedance/seedance/2.5/flash/reference-to-videoByteDance's Seedance 2.5 Flash reference-to-video model, a faster and lower-cost tier of Seedance 2.5 that generates up to 30 seconds of 480p or 720p video from reference images, videos, and audio, and can edit or extend a reference clip.
Try the model
PNG, JPEG, WebP, or GIF · 20 MiB maximum each · up to 10
What would you like to do next?
Select an example to load its inputs and result into the playground.
{
"aspect_ratio": "16:9",
"images": [
"https://static.sandbase.ai/examples/e54610d1-6a03-417d-9966-e9af7d62a688/input/b5c0e60c-f747-4424-b5f9-40ec17f196e8.jpg"
],
"prompt": "Use Image1 as the visual reference and Video1 as the motion reference. Create a peaceful cinematic shoreline scene with rolling ocean waves at golden hour, soft foam, natural water movement, and no people.",
"videos": [
"https://static.sandbase.ai/examples/e54610d1-6a03-417d-9966-e9af7d62a688/input/f6f98664-8888-482b-a862-cf4a80a6b0e8.mp4"
]
}{
"aspect_ratio": "16:9",
"images": [
"https://static.sandbase.ai/examples/5b8d9e3d-3241-42ae-bbeb-5c3a1a384119/input/64e48efa-ecbd-4a36-863c-fd1916bd25cc.jpg"
],
"prompt": "Use @Image1 as the visual reference. A slow aerial drone shot glides along this coastline at sunrise, mist over the water, gentle waves breaking on the sand."
}Seedance 2.5 Flash reference-to-video is the faster, lower-cost tier of Seedance 2.5: combine up to 10 reference images, 5 reference videos, and 5 reference audio clips to lock a character, set, motion, or voice in a new 480p or 720p clip.
| Field | Type | Default | Notes |
|---|---|---|---|
prompt | string, required | — | Up to 50,000 characters. |
images | image URL[] (≤10) | — | Reference images, Image 1, Image 2, … |
videos | video URL[] (≤5) | — | Reference clips, Video 1, Video 2, … Combined length up to 30 seconds. |
audios | audio URL[] (≤5) | — | Reference audio, Audio 1, Audio 2, … Combined length up to 30 seconds. |
task_type | auto / reference / edit / extend | auto | Reference sub-task; see below. |
duration | auto / 4–30 | 5 | Output length in seconds (4–30), or auto to let the model choose. |
aspect_ratio | 16:9 / 9:16 / 1:1 / 4:3 / 3:4 / 21:9 / adaptive | — | Output frame shape. Omit to let the model choose. |
resolution | 480p / 720p | 720p | Sets the per-second rate. |
generate_audio | boolean | true | Native dialogue, music, and effects. Same price on or off. |
seed | integer -1–4294967295 | random | Fix it to compare prompt changes; -1 is random. |
Without reference videos you pay for output seconds: $0.12 at 480p, $0.24 at 720p. With reference videos, the input video seconds (at least 2) are added to the output length and billed at a lower rate: $0.075 at 480p, $0.15 at 720p. With duration: "auto" the quote assumes 30 seconds and the final charge uses the generated length.
| Output | 480p | 720p |
|---|---|---|
| 5s, images only | $0.60 | $1.20 |
| 10s, images only | $1.20 | $2.40 |
| 5s + 5s reference video | $0.75 | $1.50 |
| 10s + 10s reference video | $1.50 | $3.00 |
Failed runs are not charged.
bytedance/seedance/2.5/reference-to-video.task_type selects the reference sub-task: reference (new video guided by the references), edit (edit a reference video), extend (extend a reference video forward or backward), or auto (default, inferred from the prompt). edit and extend need at least one video and use the source clip's aspect ratio; edit also follows the source clip's length and is billed on the actual output length. If the prompt contradicts an explicit task_type, the task fails.
{"prompt": "Replace the man in @Video1 with the character in @Image1.", "images": ["https://example.com/character.png"], "videos": ["https://example.com/source.mp4"], "task_type": "edit"}
Other Seedance 2.5 endpoints and tiers: bytedance/seedance/2.5/flash/text-to-video, bytedance/seedance/2.5/flash/image-to-video, bytedance/seedance/2.5/reference-to-video, bytedance/seedance/2.0/fast/reference-to-video.
Start building
Submit this model through the SandBase Unified Run API.
# The quoted heredoc keeps Unicode and shell metacharacters unchanged.
result=$(curl --fail-with-body --silent \
-X POST "https://api.sandbase.ai/v1/run" \
-H "Authorization: Bearer $SANDBASE_API_KEY" \
-H "Content-Type: application/json" \
--data-binary @- <<'SANDBASE_JSON'
{
"model": "bytedance/seedance/2.5/flash/reference-to-video",
"images": [
"https://static.sandbase.ai/examples/bytedance/seedance/2.0/reference-to-video/input_image_safe_ocean_0.jpg"
],
"prompt": "Use Image1 as the visual reference and Video1 as the motion reference. Create a peaceful cinematic shoreline scene with rolling ocean waves at golden hour, soft foam, natural water movement, and no people.",
"videos": [
"https://static.sandbase.ai/examples/pixverse/sound-effects/input_video_0.mp4"
],
"duration": 5,
"resolution": "720p",
"generate_audio": true
}
SANDBASE_JSON
)
run_id=$(printf '%s' "$result" | jq -r .id)
for attempt in $(seq 1 120); do
status=$(printf '%s' "$result" | jq -r .status)
case "$status" in completed|failed|timeout) break ;; esac
sleep 2
result=$(curl --fail-with-body --silent \
-H "Authorization: Bearer $SANDBASE_API_KEY" \
"https://api.sandbase.ai/v1/run/$run_id")
done
status=$(printf '%s' "$result" | jq -r .status)
[ "$status" = completed ] || { echo "Generation ended: $status" >&2; exit 1; }
printf '%s\n' "$result"Create a SandBase API key and set it in your server environment. All requests use Bearer authentication.
export SANDBASE_API_KEY="YOUR_API_KEY"Keep the key on your server. Do not expose it in browser code or commit it to a repository.
POST to /v1/run with this model ID and its inputs. The request example above includes submission and polling.
Use the returned run ID to poll until the status is completed, failed or timeout. Read output only after completed.
curl "https://api.sandbase.ai/v1/run/$RUN_ID" \
-H "Authorization: Bearer $SANDBASE_API_KEY"Pass a publicly accessible file URL in the model’s image, video or audio field. Supported formats and limits are shown beside each Playground input.
Use the upload or replace control on supported image fields in the Playground. Video URL fields accept a hosted driving video; they do not imply video-upload support.
seedintegerRandom seed for reproducible results. Omit or use -1 for a random seed.
Minimum: -1
Maximum: 4294967295
audiosarrayThe URLs of reference audio files to guide video generation. Up to 5 files; combined duration must not exceed 30 seconds.
imagesarrayThe URLs of reference images to guide video generation. Up to 10 images. Refer to them in the prompt as @Image1, @Image2, etc.
promptstringrequiredThe text prompt to generate a video from.
Default: "Use Image1 as the visual reference and Video1 as the motion reference. Create a peaceful cinematic shoreline scene with rolling ocean waves at golden hour, soft foam, natural water movement, and no people."
videosarrayThe URLs of reference videos to guide video generation. Up to 5 MP4 or MOV files; combined duration must not exceed 30 seconds.
durationvalueOutput duration in seconds (4-30), or `auto` to let the model pick a length from the prompt and inputs. `auto` is pre-authorized at 30 seconds and billed on the actual output length. When a reference video is edited or extended, use `auto` with `adaptive` aspect ratio so the output follows the source clip.
Default: 5
Possible values: "auto", 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30
task_typestring`reference` generates a new video from the references, `edit` edits a reference video (replace subjects, add or remove objects, repaint), `extend` extends a reference video forward or backward, `auto` (or omitted) lets the model infer it from the prompt. `edit` and `extend` require at least one video; `edit` forces `adaptive` aspect ratio and `auto` duration, `extend` forces `adaptive` aspect ratio. Write the prompt to match, otherwise the task can fail with a task-type mismatch.
Possible values: "auto", "reference", "edit", "extend"
resolutionstringThe resolution of the video to generate.
Default: "720p"
Possible values: "480p", "720p"
aspect_ratiostringOutput aspect ratio. Omit to let the model choose from the prompt and reference inputs.
Possible values: "16:9", "9:16", "1:1", "4:3", "3:4", "21:9", "adaptive"
generate_audiobooleanWhether to generate synchronized audio (voice, SFX, BGM) for the video.
Default: true
{
"model": "bytedance/seedance/2.5/flash/reference-to-video",
"images": [
"https://static.sandbase.ai/examples/bytedance/seedance/2.0/reference-to-video/input_image_safe_ocean_0.jpg"
],
"prompt": "Use Image1 as the visual reference and Video1 as the motion reference. Create a peaceful cinematic shoreline scene with rolling ocean waves at golden hour, soft foam, natural water movement, and no people.",
"videos": [
"https://static.sandbase.ai/examples/pixverse/sound-effects/input_video_0.mp4"
],
"duration": 5,
"resolution": "720p",
"generate_audio": true
}A completed run exposes its result in output. The following is the model’s saved example, not a new generation.
[
{
"url": "https://static.sandbase.ai/examples/e54610d1-6a03-417d-9966-e9af7d62a688/output/470160ee-d79a-4a04-a26b-eb0cef9b8a60.mp4"
}
]