> ## Documentation Index > Fetch the complete documentation index at: https://docs.openp.ai/llms.txt > Use this file to discover all available pages before exploring further. # Video generation (Seedance) > POST /seedance/v1/videos/generations — Seedance-standard video generation Generate video with Seedance models, compatible with the [seevio.ai](https://seevio.ai/api-docs) standard API. It uses an **asynchronous task** flow: submit to get a task ID, poll until done, then download the video. ## Submit a task ```bash theme={null} curl https://openp.ai/seedance/v1/videos/generations \ -H "Authorization: Bearer $OPENPAI_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "seedance-2-0-fast", "input": { "prompt": "A golden retriever running on a beach at sunset, cinematic", "duration": 5, "aspect_ratio": "16:9", "resolution": "720p" } }' ``` ### Parameters | Field | Type | Required | Description | | ------- | ------ | :------: | ----------------------------------- | | `model` | string | ✅ | See the model list below | | `input` | object | ✅ | Generation parameters, fields below | ### input fields | Field | Type | Required | Description | | ----------------- | ------- | :---------: | ------------------------------------------------------------------------------------------------------------------ | | `prompt` | string | ✅ | Video description | | `generation_type` | string | | Generation mode: `text-to-video` (default) / `image-to-video` / `reference-to-video` | | `image_urls` | array | Conditional | Image URLs. Image-to-video: 1 image (first frame) or 2 (first + last frames); reference-to-video: reference images | | `video_urls` | array | | Reference video URLs (reference-to-video only) | | `audio_urls` | array | | Reference audio URLs (reference-to-video only) | | `duration` | integer | | Duration (seconds), default 5 | | `aspect_ratio` | string | | `16:9` / `4:3` / `1:1` / `3:4` / `9:16` / `21:9` / `adaptive` (default) | | `resolution` | string | | `480p` / `720p` (default) / `1080p` | | `generate_audio` | boolean | | Generate an audio track, default `true` | | `negative_prompt` | string | | Negative prompt | Media URLs can be publicly reachable http(s) URLs or data URLs (`data:image/png;base64,...`). ### Model list `seedance-2-5` / `seedance-2-0` / `seedance-2-0-fast` / `seedance-2-0-mini` / `seedance-2-0-mini-enhanced` / `seedance-2-0-mini-enhanced` / `seedance-1-5-pro` ## Generation modes ### Image-to-video (first frame) Pass 1 image in `image_urls` as the first frame: ```bash theme={null} curl https://openp.ai/seedance/v1/videos/generations \ -H "Authorization: Bearer $OPENPAI_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "seedance-2-5", "input": { "prompt": "the subject turns toward camera, soft studio motion", "generation_type": "image-to-video", "image_urls": ["https://your.cdn.com/first-frame.jpg"], "duration": 5, "resolution": "720p" } }' ``` Media can also be inlined as data URLs (`data:;base64,...`), handy for uploading local assets without hosting them first. Image: ```json theme={null} { "model": "seedance-2-0-fast", "input": { "prompt": "gentle ocean waves under a warm sunset sky, cinematic", "generation_type": "image-to-video", "image_urls": ["data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEASABIAAD..."], "duration": 4, "resolution": "480p" } } ``` Reference videos and audios work the same way (MIME types like `video/mp4`, `audio/mpeg`): ```json theme={null} { "model": "seedance-2-5", "input": { "prompt": "the character from image 1 performs the motion from the video with the reference voice", "generation_type": "reference-to-video", "image_urls": ["data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEASABIAAD..."], "video_urls": ["data:video/mp4;base64,AAAAIGZ0eXBpc29tAAACAGlzb20..."], "audio_urls": ["data:audio/mpeg;base64,SUQzBAAAAAAAI1RTU0UAAAAPAAADTGF2ZjU4..."], "duration": 8 } } ``` Inlined video/audio inflates the request body (base64 grows \~33%); each asset is capped at 100 MB. For large or reusable assets, prefer public URLs. ### Image-to-video (first + last frames) Pass 2 images in `image_urls`; the model interpolates between the first and last frames: ```json theme={null} { "model": "seedance-2-5", "input": { "prompt": "smooth camera push-in from the first frame to the last", "generation_type": "image-to-video", "image_urls": [ "https://your.cdn.com/first-frame.jpg", "https://your.cdn.com/last-frame.jpg" ], "duration": 5 } } ``` ### Reference-to-video (image + video) Lock the subject with a reference image and replicate camera motion from a reference video: ```json theme={null} { "model": "seedance-2-5", "input": { "prompt": "use the product from image 1 and the camera motion from the reference video", "generation_type": "reference-to-video", "image_urls": ["https://your.cdn.com/product.jpg"], "video_urls": ["https://your.cdn.com/camera-motion.mp4"], "duration": 8, "resolution": "720p" } } ``` ### Reference-to-video (image + audio) Make the subject speak or sing along to a reference audio: ```json theme={null} { "model": "seedance-2-5", "input": { "prompt": "the woman from image 1 sings the reference song on stage", "generation_type": "reference-to-video", "image_urls": ["https://your.cdn.com/singer.jpg"], "audio_urls": ["https://your.cdn.com/song.mp3"], "duration": 10 } } ``` ### Reference-to-video (image + video + audio) All three material types can be combined: ```json theme={null} { "model": "seedance-2-5", "input": { "prompt": "the character from image 1 performs the motion from the video with the reference voice", "generation_type": "reference-to-video", "image_urls": ["https://your.cdn.com/character.jpg"], "video_urls": ["https://your.cdn.com/motion.mp4"], "audio_urls": ["https://your.cdn.com/voice.mp3"], "duration": 8 } } ``` ### Reference-to-video (audio only) `seedance-2-5` accepts audio as the only reference; the `seedance-2-0` family requires at least one image or video alongside audio: ```json theme={null} { "model": "seedance-2-5", "input": { "prompt": "a music video visualizer following the rhythm of the reference track", "generation_type": "reference-to-video", "audio_urls": ["https://your.cdn.com/track.mp3"], "duration": 10 } } ``` ### Material limits | Model | Images | Videos | Audios | Total | | --------------------- | ------ | -------------------------------- | -------------------------------- | ----- | | `seedance-2-5` | ≤ 30 | ≤ 10, each 2-30s, combined ≤ 30s | ≤ 10, each 2-30s, combined ≤ 30s | ≤ 50 | | `seedance-2-0` family | ≤ 9 | ≤ 3, combined ≤ 15s | ≤ 3, combined ≤ 15s | — | ## Response (task created) ```json theme={null} { "taskId": "task_id", "credits": 28 } ``` `taskId` is the task ID; `credits` is the estimated charge. ## Query task status ```bash theme={null} curl https://openp.ai/seedance/v1/tasks/task_id \ -H "Authorization: Bearer $OPENPAI_API_KEY" ``` ```json theme={null} { "id": "task_id", "status": "completed", "created_at": 1786176476, "model": "seedance-2-0-fast", "billing_status": "charged", "credits": 28, "data": { "results": [ "https://openp.ai/seedance/v1/videos/generations/task_id.mp4" ], "processing_time": 104 } } ``` | Field | Description | | ---------------------- | ----------------------------------------------------------- | | `status` | `queued` / `generating` / `completed` / `failed` | | `billing_status` | `reserved` (pre-charged) / `charged` / `refunded` | | `credits` | Actual charge | | `failed_reason` | Failure reason (only when `failed`; no `data` in that case) | | `data.results` | Video download URLs (only when `completed`) | | `data.processing_time` | Generation time (seconds) | ## Download the video ```bash theme={null} curl -L https://openp.ai/seedance/v1/videos/generations/task_id.mp4 \ -H "Authorization: Bearer $OPENPAI_API_KEY" \ -o video.mp4 ``` Returns a `video/mp4` byte stream. ## Polling advice * Wait 5-10 seconds before the first poll. * Poll every 10-30 seconds. * Tasks typically finish in 1-3 minutes. * On failure, `failed_reason` explains why. ## Billing Billed as **model unit price × seconds**, pre-charged at submit and settled on completion. Failed tasks are refunded automatically (`billing_status` becomes `refunded`).