Skip to main content
Unifically LogoUnificAlly
Model logo

SeeDance 1.0 Fast API

    PlaygroundREADME
    SeeDance 1.0SeeDance 1.0 MiniSeeDance 1.0 Fast
    Output

    Your generated video will appear here

    SeeDance 1.0 Fast

    Unifically: one endpoint, one API key, USD billing.

    What is SeeDance 1.0 Fast?

    SeeDance 1.0 Fast is the throughput-focused variant of SeeDance 1.0. It adds two things on top of the base 1.0 Pro model: 1080p output for single-shot modes, and a multi_frame mode that chains multiple keyframe segments into one continuous clip. Multi-frame mode is the differentiator: you pass a frames array where each entry has a per-frame prompt, an optional keyframe image_url, and a per-segment duration, and the model stitches them into a single output. Single-shot modes (text_to_video and first_last_frame) honor a 5 or 10 second duration at 720p or 1080p. An integer seed makes any of the runs reproducible.

    Key features of SeeDance 1.0 Fast

    Four features cover what teams build with 1.0 Fast.

    Multi-frame chained segments in one call

    Pass a frames array with per-segment prompts, keyframe images, and durations. The model stitches the segments into one continuous clip. Useful for multi-beat ads and sequence storytelling without manual stitching.

    1080p output for single-shot modes

    text_to_video and first_last_frame both honor a resolution choice of 720p or 1080p. Use 1080p for hero clips that need web-quality delivery.

    All three SeeDance 1.0 modes on one model

    text_to_video, first_last_frame, and multi_frame share the same model ID. The mode field picks which path runs and which fields apply.

    Integer seed for reproducible runs

    Pass a seed alongside the prompt and frames. Re-sending the same inputs returns the same result, which is useful for locking a chosen take or running regression tests.

    Best for

    Multi-beat ad sequences

    Chained keyframe segments with their own prompts. Useful for narrative ads with a clear arc across the cut.

    Storyboard-to-animatic

    Each storyboard frame becomes a segment. The model produces motion between them in one call.

    Long-form social

    Chain several 10-second segments into one continuous clip without an NLE step.

    1080p single-shot work

    Higher resolution than 1.0 Pro for hero clips that need web-quality delivery.

    Throughput-heavy pipelines

    Optimized for batch jobs. Useful when you queue many concurrent generations.

    Variants

    SeeDance 1.0 Fast has three modes selected by the mode field, plus a 720p / 1080p resolution choice. Pick by what the call needs.

    text_to_video

    Prompt-only path. You control the aspect ratio across six options (1:1, 3:4, 4:3, 9:16, 16:9, 21:9). 5 or 10 seconds at 720p or 1080p.

    first_last_frame

    Pass first_frame_url (required) and optionally last_frame_url. The model interpolates the motion between them. Aspect ratio is auto-detected from the input image.

    multi_frame

    Pass a frames array with per-segment prompts, optional keyframe images, and per-segment durations. The model stitches the segments into one clip. Total length is the sum of per-frame durations.

    Use cases

    Build a multi-beat product ad as one continuous clip by writing a frames array with three segments (hero reveal, feature highlight, closing card) and letting the model stitch them. Convert a storyboard into a working animatic by feeding each board as a keyframe in the frames array with the matching shot description as the prompt. Render a 1080p hero clip from key art by running text_to_video at 1080p with a 9:16 aspect for vertical placements. Lock a chosen take across an iteration cycle by passing the same seed on every run.

    API examples

    Call SeeDance 1.0 Fast from any language by POSTing to /v1/tasks. Full parameter docs live at docs.unifically.com/models/video/bytedance/seedance-1.0-fast.

    curl -X POST https://api.unifically.com/v1/tasks \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "bytedance/seedance-1.0-fast",
    "input": {
      "prompt": "A person breakdancing in an urban setting",
      "mode": "text_to_video",
      "aspect_ratio": "9:16",
      "duration": 10,
      "resolution": "1080p"
    }
  }'

    Successful submission returns a task_id. Poll GET /v1/tasks/<task_id> or set a callback_url on the request to receive the finished result.

    FAQs

    People also ask

    SeeDance 1.0 Fast is the throughput-focused variant of SeeDance 1.0. It adds 1080p output and multi-frame mode (chain multiple keyframe segments into one continuous clip) on top of the standard text-to-video and first-and-last-frame modes.

    You author a frames array. Each frame entry has a per-frame prompt, an optional keyframe image_url, and a per-segment duration in seconds. The model stitches them into one continuous output. At least one frame must include an image_url, and total clip length is the sum of all frame durations.

    Same parameter shape and the same three modes (text_to_video, first_last_frame, multi_frame). Different billing key and performance profile. Pick based on your budget vs latency needs and your team's existing pipeline.

    5 or 10 seconds for text_to_video and first_last_frame. In multi_frame mode total length is the sum of per-segment durations, so you can chain longer clips.

    No. SeeDance 1.0 Fast is video-only. Use SeeDance 2.0 if you need native audio in the same call.