> ## Documentation Index
> Fetch the complete documentation index at: https://docs.unifically.com/llms.txt
> Use this file to discover all available pages before exploring further.

# SeeDance 2.5

> Long-form multimodal video generation with native audio

Generate videos using ByteDance SeeDance 2.5 with text-to-video, first/last frame, and multimodal reference workflows. It supports image, video, and audio references, synchronized native audio, 4–30 second durations, and 480p or 720p output.

## Model

```
bytedance/seedance-2.5
```

<Note>
  Unifically automatically routes prompt-only requests through the text generation channel and requests containing any reference media through the multimodal channel. No routing option is required in your request.
</Note>

## Request types

| Request type         | Required inputs                      | Optional inputs                                                                                              | Notes                                                                                                      |
| -------------------- | ------------------------------------ | ------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------- |
| Text-to-video        | `prompt`                             | `aspect_ratio`, `duration`, `resolution`, `generate_audio`, `seed`                                           | Do not include reference media URLs                                                                        |
| First/last frame     | `first_frame_url`, `last_frame_url`  | `prompt`, `duration`, `resolution`, `generate_audio`, `seed`                                                 | Both frame URLs are required. `aspect_ratio` must be `adaptive` because the output follows the first frame |
| Multimodal reference | `prompt`, at least one reference URL | `image_urls`, `video_urls`, `audio_urls`, `aspect_ratio`, `duration`, `resolution`, `generate_audio`, `seed` | Reference assets in the prompt as `[Image1]`, `[Video1]`, and `[Audio1]`                                   |

## Parameters

| Parameter         | Type      | Required    | Default      | Description                                                                                                                                                                                                                   |
| ----------------- | --------- | ----------- | ------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `prompt`          | string    | Conditional | `""`         | Creative description (max 5,000 characters). Required for text-to-video and multimodal reference requests; optional for first/last frame requests. Reference media in array order using `[Image1]`, `[Video1]`, or `[Audio1]` |
| `resolution`      | string    | No          | `"480p"`     | `480p` or `720p`                                                                                                                                                                                                              |
| `duration`        | integer   | No          | `4`          | Output duration in whole seconds (4–30)                                                                                                                                                                                       |
| `aspect_ratio`    | string    | No          | `"adaptive"` | `adaptive`, `1:1`, `3:4`, `4:3`, `9:16`, `16:9`, or `21:9`. Must be `adaptive` for first/last frame requests                                                                                                                  |
| `generate_audio`  | boolean   | No          | `true`       | Generate synchronized voice, sound effects, and background music with the video                                                                                                                                               |
| `seed`            | integer   | No          | Random       | Seed from `-1` to `2147483647`. Use the same numeric seed for reproducible results; `-1` selects a random seed                                                                                                                |
| `first_frame_url` | string    | Conditional | `null`       | First frame image URL. Must be provided together with `last_frame_url` and cannot be combined with reference arrays                                                                                                           |
| `last_frame_url`  | string    | Conditional | `null`       | Last frame image URL. Must be provided together with `first_frame_url` and cannot be combined with reference arrays                                                                                                           |
| `image_urls`      | string\[] | No          | `null`       | Up to 30 reference image URLs. Reference them as `[Image1]`, `[Image2]`, and so on                                                                                                                                            |
| `video_urls`      | string\[] | No          | `null`       | Up to 10 reference video URLs with a combined duration of up to 30 seconds. Each video must contain at least 409,600 pixels. Reference them as `[Video1]`, `[Video2]`, and so on                                              |
| `audio_urls`      | string\[] | No          | `null`       | Up to 10 reference audio URLs with a combined duration of up to 30 seconds. Audio-only reference requests are supported. Reference them as `[Audio1]`, `[Audio2]`, and so on                                                  |

The total number of reference assets across images, videos, and audio cannot exceed 50. First/last frame URLs cannot be combined with `image_urls`, `video_urls`, or `audio_urls`.

## Example - Text-to-Video

```bash theme={null}
curl -X POST https://api.unifically.com/v1/tasks \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "bytedance/seedance-2.5",
    "input": {
      "prompt": "A cinematic tracking shot through a bustling night market in the rain, with natural crowd ambience and distant music",
      "aspect_ratio": "16:9",
      "duration": 12,
      "resolution": "720p",
      "generate_audio": true
    }
  }'
```

## Example - First & Last Frame

```bash theme={null}
curl -X POST https://api.unifically.com/v1/tasks \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "bytedance/seedance-2.5",
    "input": {
      "prompt": "The camera glides forward as sunrise gradually illuminates the valley",
      "duration": 10,
      "resolution": "720p",
      "first_frame_url": "https://example.com/start.jpg",
      "last_frame_url": "https://example.com/end.jpg"
    }
  }'
```

## Example - Video and Audio References

```bash theme={null}
curl -X POST https://api.unifically.com/v1/tasks \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "bytedance/seedance-2.5",
    "input": {
      "prompt": "Create a new cinematic sequence inspired by the camera movement in [Video1] and timed to the rhythm of [Audio1]",
      "duration": 15,
      "resolution": "720p",
      "video_urls": [
        "https://example.com/camera-reference.mp4"
      ],
      "audio_urls": [
        "https://example.com/music-reference.mp3"
      ]
    }
  }'
```

References are numbered independently and in array order: the first image is `[Image1]`, the first video is `[Video1]`, and the first audio clip is `[Audio1]`.

## Response

```json theme={null}
{
  "code": 200,
  "success": true,
  "data": {
    "task_id": "abc123def456",
    "status": "processing"
  }
}
```