---
title: Seedance 2.5 | Runware Docs
url: https://runware.ai/docs/models/bytedance-seedance-2-5
description: Professional multimodal video generation with native 30-second clips, large-scale reference control, and precise localized editing
---
# Seedance 2.5

![Seedance 2.5](https://assets.runware.ai/covers/bytedance-seedance-2-5.jpg)

Seedance 2.5 is ByteDance's higher-end multimodal video generation model for production-oriented creative work. It supports native 30-second video generation, larger reference sets across image, video, and audio inputs, and stronger instruction following for complex shots, multilingual prompts, and reference-heavy creative direction. It is especially well suited to branded storytelling, advertising, explainers, and other commercial workflows that need longer continuous scenes, tighter subject and style consistency, and precise edits that preserve the rest of the shot.

- **Status**: live
- **ID**: `bytedance:seedance@2.5`
- Try in Playground
- **Creator**: ByteDance
- **Release Date**: August 7, 2026
- **Capabilities**: Text to Video, Image to Video, Video to Video, Audio to Video, Edit, Checkpoint, Audio

## Pricing

$0.102 from, per second · ~2m 56s per request

**Rates** (per second)

*Text/Image to Video*

- 480p: `$0.102`
- 720p: `$0.23`
- 1080p: `$0.614`
- 4K: `$0.921`

*Video-to-Video*

- 480p: `$0.131`
- 720p: `$0.295`
- 1080p: `$0.734`
- 4K: `$1.14`

**6 example generations**

- [text-to-video](https://runware.ai/docs/models/bytedance-seedance-2-5/examples#example-midwifery-archive-restoration-reveal) · 720p · 5s · audio: `$1.17` (~2m 24s)
- [video-edit](https://runware.ai/docs/models/bytedance-seedance-2-5/examples#example-wildlife-tracking-collar-cleanup) · 720p · autos: `$1.78` (~2m 56s)
- [reference-to-video](https://runware.ai/docs/models/bytedance-seedance-2-5/examples#example-rooftop-apiary-scene-replacement) · 720p · 8s: `$1.86` (~3m 53s)
- [first-last-frame](https://runware.ai/docs/models/bytedance-seedance-2-5/examples#example-four-track-recorder-colorway-reveal-film) · 720p · 10s: `$2.32` (~1m 45s)
- [first-frame](https://runware.ai/docs/models/bytedance-seedance-2-5/examples#example-geothermal-spa-equipment-cleanup) · 1080p · 6s: `$3.67` (~2m 56s)
- [video-extend](https://runware.ai/docs/models/bytedance-seedance-2-5/examples#example-quarry-photo-linocut-transition) · 1080p · 6s: `$4.37` (~3m 22s)

times vary, measured end to end · [How pricing works](https://runware.ai/docs/models-api/pricing)

## Compatibility & Validation

Inside `inputs`, `frameImages` cannot be used with `referenceImages`, `referenceVideos`, `referenceAudios`, or `video`.

---

When `inputs.draftId` is provided, `positivePrompt`, `duration`, `seed`, `settings.audio`, `settings.operation`, `inputs.video`, `inputs.frameImages`, `inputs.referenceImages`, `inputs.referenceVideos`, and `inputs.referenceAudios` cannot be used.

---

When `inputs.draftId` is provided, `resolution` is limited to `1080p`, `4k`.

---

When `inputs.draftId` is provided, dimensions are limited to `1920×1080`, `1664×1248`, `1440×1440`, `1248×1664`, `1080×1920`, `2206×946`, `3840×2160`, `3326×2494`, `2880×2880`, `2494×3326`, `2160×3840` or `4398×1886`.

---

When `inputs.frameImages` or `inputs.video` is provided, `width/height` cannot be used.

---

When none of `inputs.frameImages`, `inputs.referenceImages`, `inputs.referenceVideos`, `inputs.referenceAudios`, `inputs.video`, or `inputs.draftId` are provided, `positivePrompt` is required.

---

`resolution` cannot be used with `width/height`.

---

`settings.draft` cannot be used with `inputs.draftId`.

---

When `settings.draft` is `true`, `resolution` must be `480p`.

---

When `settings.draft` is `true`, dimensions are limited to `854×480`, `752×560`, `640×640`, `560×752`, `480×854` or `992×432`.

---

When `settings.operation` is provided, `inputs.video` is required.

---

When `settings.operation` is `edit`, `duration` must be `auto`.

---

`width` and `height` must be used together.

---

The following dimension combinations are supported:

| Configuration | Dimensions |
| --- | --- |
| `480p (~16:9)` | `854x480` |
| `480p (~4:3)` | `752x560` |
| `480p (1:1)` | `640x640` |
| `480p (~3:4)` | `560x752` |
| `480p (~9:16)` | `480x854` |
| `480p (~21:9)` | `992x432` |
| `720p (16:9)` | `1280x720` |
| `720p (4:3)` | `1112x834` |
| `720p (1:1)` | `960x960` |
| `720p (3:4)` | `834x1112` |
| `720p (9:16)` | `720x1280` |
| `720p (21:9)` | `1470x630` |
| `1080p (16:9)` | `1920x1080` |
| `1080p (4:3)` | `1664x1248` |
| `1080p (1:1)` | `1440x1440` |
| `1080p (3:4)` | `1248x1664` |
| `1080p (9:16)` | `1080x1920` |
| `1080p (~21:9)` | `2206x946` |
| `4K (16:9)` | `3840x2160` |
| `4K (~4:3)` | `3326x2494` |
| `4K (1:1)` | `2880x2880` |
| `4K (~3:4)` | `2494x3326` |
| `4K (9:16)` | `2160x3840` |
| `4K (~21:9)` | `4398x1886` |

## Notes

Seedance 2.5 supports multiple generation modes:

- **Text-to-Video:** Provide a prompt on its own. Add `width`/`height` or `resolution` to pin the output, or omit both and the model picks the aspect ratio from the prompt.
- **Image-to-Video:** Provide frame images via `inputs.frameImages` (up to 2, first/last). Cannot be combined with reference images or reference videos.
- **Reference-Guided:** Provide reference images (`inputs.referenceImages`, up to 30), reference videos (`inputs.referenceVideos`, up to 10), and reference audios (`inputs.referenceAudios`, up to 10). Any combination works, and audio can be used on its own.
- **Editing and Extension:** Provide the source clip via `inputs.video`. Describe the change in the prompt to edit it, or ask for a continuation to extend it. Set `settings.operation` to `edit` or `extend` to choose one yourself.
- **Draft and Final Render:** Set `settings.draft` for a fast 480p preview, which returns a `draftId`. Send it as `inputs.draftId` to render that preview at 1080p or 4K. The prompt, media, duration and seed come from the draft and cannot be sent again.

A prompt is required for text-to-video. The other modes that take media accept one but do not need it.

**Sizing:** Use `width`/`height` for an explicit aspect ratio pair, or `resolution` to set the tier and let the model pick the ratio. The two cannot be combined, and both are optional. Frame images pin the aspect ratio to the first frame, so `width`/`height` cannot be sent with them.

`duration` takes a whole number of seconds, or `auto` to let the model choose.

Editing and extension both send the source clip as `inputs.video`. When `settings.operation` is omitted, the prompt decides which one happens. Both inherit the aspect ratio of that clip, so `width`/`height` cannot be sent with it, though `resolution` still selects the tier. They differ on length: extension accepts a number of seconds or `auto`, while editing accepts only `auto`, since the output matches the length of the source clip.

**Reference videos are guidance, not a clip to change.** When the prompt asks to edit a reference video, Seedance 2.5 treats the request as an edit, and an edit takes its aspect ratio and length from that video. Sent with `width`/`height` or a numeric `duration`, such a request is accepted and then fails during generation. To edit a clip, send it as `inputs.video` with `settings.operation` set to `edit`, so the same mistake is rejected when the request is submitted. To keep the clip in `inputs.referenceVideos`, omit `width`/`height` and send `duration: "auto"`.

## Request Parameters

**API Options**

Platform-level options for task execution and delivery.

### [taskType](https://runware.ai/docs/models/bytedance-seedance-2-5#request-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `videoInference`

Identifier for the type of task being performed.

### [taskUUID](https://runware.ai/docs/models/bytedance-seedance-2-5#request-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier for tracking tasks and matching async responses. Must be unique per task.

### [outputType](https://runware.ai/docs/models/bytedance-seedance-2-5#request-outputtype)

- **Type**: `string`
- **Default**: `URL`

Video output type.

**Allowed values**: `URL`

### [outputFormat](https://runware.ai/docs/models/bytedance-seedance-2-5#request-outputformat)

- **Type**: `string`
- **Default**: `MP4`

Specifies the file format of the generated output. The available values depend on the task type and the specific model's capabilities.

- \`MP4\`: Widely supported video container (H.264), recommended for general use.
- \`WEBM\`: Optimized for web delivery.
- \`MOV\`: QuickTime format, common in professional workflows (Apple ecosystem).

**Allowed values**: `MP4` `WEBM` `MOV`

### [outputQuality](https://runware.ai/docs/models/bytedance-seedance-2-5#request-outputquality)

- **Type**: `integer`
- **Min**: `20`
- **Max**: `99`
- **Default**: `95`

Compression quality of the output. Higher values preserve quality but increase file size.

### [webhookURL](https://runware.ai/docs/models/bytedance-seedance-2-5#request-webhookurl)

- **Type**: `string`
- **Format**: `uri`

Webhook URL that receives JSON responses via HTTP POST when generation tasks complete. For batch requests with multiple results, each completed item triggers a separate webhook call as it becomes available.

**Learn more** (1 resource):

- [Webhooks](https://runware.ai/docs/models-api/webhooks) (platform)

### [deliveryMethod](https://runware.ai/docs/models/bytedance-seedance-2-5#request-deliverymethod)

- **Type**: `string`
- **Default**: `async`

Determines how the API delivers task results.

**Allowed values**:

- `sync` Returns results directly in the API response. Over HTTP, a task still running after 120 seconds continues as an async task and comes back with status 'processing'. Poll for its results using getResponse.
- `async` Returns an immediate acknowledgment with the task UUID. Poll for results using getResponse.

**Learn more** (1 resource):

- [Task Polling](https://runware.ai/docs/models-api/task-polling) (platform)

### [uploadEndpoint](https://runware.ai/docs/models/bytedance-seedance-2-5#request-uploadendpoint)

- **Type**: `string`
- **Format**: `uri`

Specifies a URL where the generated content will be automatically uploaded using the HTTP PUT method. The raw binary data of the media file is sent directly as the request body. For secure uploads to cloud storage, use presigned URLs that include temporary authentication credentials.

**Common use cases:**

- **Cloud storage**: Upload directly to S3 buckets, Google Cloud Storage, or Azure Blob Storage using presigned URLs.
- **CDN integration**: Upload to content delivery networks for immediate distribution.

```text
// S3 presigned URL for secure upload
https://your-bucket.s3.amazonaws.com/generated/content.mp4?X-Amz-Signature=abc123&X-Amz-Expires=3600

// Google Cloud Storage presigned URL
https://storage.googleapis.com/your-bucket/content.jpg?X-Goog-Signature=xyz789

// Custom storage endpoint
https://storage.example.com/uploads/generated-image.jpg
```

The content data will be sent as the request body to the specified URL when generation is complete.

### [safety](https://runware.ai/docs/models/bytedance-seedance-2-5#request-safety)

- **Path**: `safety.checkContent`
- **Type**: `object (2 properties)`

Content safety checking configuration for video generation.

#### [checkContent](https://runware.ai/docs/models/bytedance-seedance-2-5#request-safety-checkcontent)

- **Path**: `safety.checkContent`
- **Type**: `boolean`

Enable or disable content safety checking. Increases total generation time and adds a small charge on top of the generation cost.

#### [mode](https://runware.ai/docs/models/bytedance-seedance-2-5#request-safety-mode)

- **Path**: `safety.mode`
- **Type**: `string`
- **Default**: `fast`

Safety checking mode for video generation.

**Allowed values**:

- `fast` Checks key frames.
- `full` Checks all frames.

### [ttl](https://runware.ai/docs/models/bytedance-seedance-2-5#request-ttl)

- **Type**: `integer`
- **Min**: `60`

Time-to-live (TTL) in seconds for generated content. Only applies when `outputType` is `URL`.

### [includeCost](https://runware.ai/docs/models/bytedance-seedance-2-5#request-includecost)

- **Type**: `boolean`

Include task cost in the response.

### [numberResults](https://runware.ai/docs/models/bytedance-seedance-2-5#request-numberresults)

- **Type**: `integer`
- **Min**: `1`
- **Max**: `4`
- **Default**: `1`

Number of results to generate. Each result uses a different seed, producing variations of the same parameters.

**Inputs**

Input resources for the task (images, audio, etc). These must be nested inside the \`inputs\` object.

### [referenceImages](https://runware.ai/docs/models/bytedance-seedance-2-5#request-inputs-referenceimages)

- **Path**: `inputs.referenceImages`
- **Type**: `array of strings`
- **Min items**: `1`
- **Max items**: `30`
- **File size (each)**: `up to 30 MB`
- **Size (each)**: `300 px to 6000 px`
- **Aspect ratio (each)**: `1:2.5 to 2.5:1`

List of reference images (UUID, URL, Data URI, or Base64).

### [frameImages](https://runware.ai/docs/models/bytedance-seedance-2-5#request-inputs-frameimages)

- **Path**: `inputs.frameImages`
- **Type**: `array of strings or objects`
- **Min items**: `1`
- **Max items**: `2`
- **Size (each)**: `300 px to 6000 px`
- **Aspect ratio (each)**: `1:2.5 to 2.5:1`

An array of frame-specific image inputs to guide video generation. Each item can be either a plain image input (UUID, URL, Data URI, or Base64) or an object that pairs an image with a target position in the video.

The `frameImages` parameter allows you to constrain specific frames within the video sequence, ensuring that particular visual content appears at designated points. Position can be specified using `frame` (named positions or frame indices) or `timestamp` (seconds), depending on model support. This is different from `referenceImages`, which provide overall visual guidance without constraining specific timeline positions.

When the `frame` parameter is omitted, automatic distribution rules apply:

- **1 image**: Used as the first frame.
- **2 images**: First and last frames.

**Examples**:

**Shorthand format:** When you don't need to specify a frame position, you can pass a plain image input directly.

```json
"frameImages": [
  "aac49721-1964-481a-ae78-8a4e29b91402"
]
```

**Object format:** When you need to specify a position, use an object with `image` and either `frame` or `timestamp` (model-dependent).

```json
"frameImages": [
  {
    "image": "aac49721-1964-481a-ae78-8a4e29b91402",
    "frame": "first"
  }
]
```

**First and last frames:** With two images, they automatically become the first and last frames of the video sequence. You can mix shorthand and object formats.

```json
"frameImages": [
  "aac49721-1964-481a-ae78-8a4e29b91402",
  {
    "image": "3ad204c3-a9de-4963-8a1a-c3911e3afafe",
    "frame": "last"
  }
]
```

**Format 1: string[]**:

- **Type**: `string`

Image input (UUID, URL, Data URI, or Base64).

**Format 2: object[]**:

#### [image](https://runware.ai/docs/models/bytedance-seedance-2-5#request-inputs-frameimages-format-2-image)

- **Path**: `inputs.frameImages.image`
- **Type**: `string`
- **Required**: true

Image input (UUID, URL, Data URI, or Base64).

#### [frame](https://runware.ai/docs/models/bytedance-seedance-2-5#request-inputs-frameimages-format-2-frame)

- **Path**: `inputs.frameImages.frame`
- **Type**: `object`

Target frame position for the image. Supports first and last frame.

**Allowed values**:

- `first` First frame of the video.
- `last` Last frame of the video.
- `0` Frame index 0 (first frame).
- `-1` Frame index -1 (last frame).

### [video](https://runware.ai/docs/models/bytedance-seedance-2-5#request-inputs-video)

- **Path**: `inputs.video`
- **Type**: `string`
- **File size**: `up to 200 MB`
- **Duration**: `2s to 30s`
- **Size**: `300 px to 6000 px`
- **Aspect ratio**: `1:2.5 to 2.5:1`

Video input (UUID or URL). The source clip to edit or extend.

### [referenceVideos](https://runware.ai/docs/models/bytedance-seedance-2-5#request-inputs-referencevideos)

- **Path**: `inputs.referenceVideos`
- **Type**: `array of strings`
- **Min items**: `1`
- **Max items**: `10`
- **File size (each)**: `up to 200 MB`
- **Duration (each)**: `2s to 30s`
- **Duration (total)**: `up to 30s`
- **Size (each)**: `300 px to 6000 px`
- **Aspect ratio (each)**: `1:2.5 to 2.5:1`

List of reference videos (UUID or URL). Used as visual guidance for a new video. To edit or extend an existing clip, use `inputs.video` instead.

### [referenceAudios](https://runware.ai/docs/models/bytedance-seedance-2-5#request-inputs-referenceaudios)

- **Path**: `inputs.referenceAudios`
- **Type**: `array of strings`
- **Min items**: `1`
- **Max items**: `10`
- **File size (each)**: `up to 15 MB`
- **Duration (each)**: `2s to 30s`
- **Duration (total)**: `up to 30s`

List of reference audios (UUID or URL).

### [draftId](https://runware.ai/docs/models/bytedance-seedance-2-5#request-inputs-draftid)

- **Path**: `inputs.draftId`
- **Type**: `string`
- **Min**: `1`

Draft ID from a previous draft mode generation. Provide this to render that 480p preview as the final video. Prompt and media come from the draft and cannot be sent again.

**Core Parameters**

Primary parameters that define the task output.

### [model](https://runware.ai/docs/models/bytedance-seedance-2-5#request-model)

- **Type**: `string`
- **Required**: true
- **Value**: `bytedance:seedance@2.5`

Send `auto` instead of an AIR identifier to have the model router choose the model for this task. The rest of the request is unchanged, every value you send is kept, and only the parameters you leave out are filled in. See [Model Router](https://runware.ai/docs/models-api/model-router).

**Learn more** (3 resources):

- [Text To Image: Model Selection](https://runware.ai/docs/learn/text-to-image#model-selection) (learn)
- [Image Inpainting: Model Specialized Inpainting Models](https://runware.ai/docs/learn/image-inpainting#model-specialized-inpainting-models) (learn)
- [Image Outpainting: Other Critical Parameters](https://runware.ai/docs/learn/image-outpainting#other-critical-parameters) (learn)

### [positivePrompt](https://runware.ai/docs/models/bytedance-seedance-2-5#request-positiveprompt)

- **Type**: `string`
- **Required**: conditional — Required when inputs.frameImages and inputs.referenceImages and inputs.referenceVideos and inputs.referenceAudios and inputs.video and inputs.draftId are not provided
- **Min**: `2`
- **Max**: `50000`

Text prompt describing elements to include in the generated output. Reference assets in array order as `@Image1`, `@Video1`, and `@Audio1`, numbered from one.

**Learn more** (1 resource):

- [Prompts](https://runware.ai/docs/learn/prompts) (learn)

### [width](https://runware.ai/docs/models/bytedance-seedance-2-5#request-width)

- **Type**: `integer`
- **Paired with**: height

Width of the generated media in pixels.

**Learn more** (2 resources):

- [Dimensions](https://runware.ai/docs/learn/dimensions) (learn)
- [Image Outpainting: Dimensions Critical For Outpainting](https://runware.ai/docs/learn/image-outpainting#dimensions-critical-for-outpainting) (learn)

### [height](https://runware.ai/docs/models/bytedance-seedance-2-5#request-height)

- **Type**: `integer`
- **Paired with**: width

Height of the generated media in pixels.

**Learn more** (2 resources):

- [Dimensions](https://runware.ai/docs/learn/dimensions) (learn)
- [Image Outpainting: Dimensions Critical For Outpainting](https://runware.ai/docs/learn/image-outpainting#dimensions-critical-for-outpainting) (learn)

### [resolution](https://runware.ai/docs/models/bytedance-seedance-2-5#request-resolution)

- **Type**: `string`
- **Default**: `720p`

Resolution preset for the output. When used with input media, automatically matches the aspect ratio from the input.

**Allowed values**: `480p` `720p` `1080p` `4k`

### [duration](https://runware.ai/docs/models/bytedance-seedance-2-5#request-duration)

- **Type**: `integer | string`
- **Min**: `4`
- **Max**: `30`
- **Default**: `5`

Length of the generated video in seconds. The total number of frames produced is determined by duration multiplied by the model's frame rate (fps).

**Allowed values**: `auto`

### [seed](https://runware.ai/docs/models/bytedance-seedance-2-5#request-seed)

- **Type**: `integer`
- **Min**: `0`
- **Max**: `2147483647`

Random seed for reproducible generation. When not provided, a random seed is generated in the unsigned 32-bit range.

**Settings**

Technical parameters to fine-tune the inference process. These must be nested inside the \`settings\` object.

### [audio](https://runware.ai/docs/models/bytedance-seedance-2-5#request-settings-audio)

- **Path**: `settings.audio`
- **Type**: `boolean`
- **Default**: `true`

Generate synchronized audio.

### [draft](https://runware.ai/docs/models/bytedance-seedance-2-5#request-settings-draft)

- **Path**: `settings.draft`
- **Type**: `boolean`

Enable draft mode for rapid iteration. Generates a fast 480p preview. Use the returned draft ID with `inputs.draftId` to produce full quality output from an approved draft.

### [operation](https://runware.ai/docs/models/bytedance-seedance-2-5#request-settings-operation)

- **Path**: `settings.operation`
- **Type**: `string`

Operation to perform on the source video. When omitted, the prompt decides which one runs.

**Allowed values**:

- `edit` Modify the source clip.
- `extend` Continue the source clip beyond its original ending.

## Response Parameters

### [taskType](https://runware.ai/docs/models/bytedance-seedance-2-5#response-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `videoInference`

Identifier for the type of task this response belongs to.

### [taskUUID](https://runware.ai/docs/models/bytedance-seedance-2-5#response-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier echoed from the original request, used to match async responses to their tasks.

### [videoUUID](https://runware.ai/docs/models/bytedance-seedance-2-5#response-videouuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID of the output video.

### [videoURL](https://runware.ai/docs/models/bytedance-seedance-2-5#response-videourl)

- **Type**: `string`
- **Format**: `uri`

URL of the output video.

### [videoBase64Data](https://runware.ai/docs/models/bytedance-seedance-2-5#response-videobase64data)

- **Type**: `string`

Base64-encoded video data.

### [videoDataURI](https://runware.ai/docs/models/bytedance-seedance-2-5#response-videodatauri)

- **Type**: `string`
- **Format**: `uri`

Data URI of the output video.

### [seed](https://runware.ai/docs/models/bytedance-seedance-2-5#response-seed)

- **Type**: `integer`

The seed used for generation. If none was provided, shows the randomly generated seed.

### [NSFWContent](https://runware.ai/docs/models/bytedance-seedance-2-5#response-nsfwcontent)

- **Type**: `boolean`

Flag indicating if NSFW content was detected.

### [cost](https://runware.ai/docs/models/bytedance-seedance-2-5#response-cost)

- **Type**: `float`

Task cost in USD. Present when `includeCost` is set to `true` in the request.

### [outputs](https://runware.ai/docs/models/bytedance-seedance-2-5#response-outputs)

- **Path**: `outputs.draftId`
- **Type**: `object (1 property)`

Generated output artifacts.

#### [draftId](https://runware.ai/docs/models/bytedance-seedance-2-5#response-outputs-draftid)

- **Path**: `outputs.draftId`
- **Type**: `string`

Identifier for the generated draft, returned when draft mode is enabled. Provide it as `inputs.draftId` to render the approved preview at full quality.

## Examples

### text-to-video (text-to-video)

[Watch video](https://assets.runware.ai/examples/bytedance-seedance-2-5/e7910c98-6bff-4315-92f5-7f12f1a34f66.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "4426ac8c-3cfe-4c43-90a1-14db36341627",
  "model": "bytedance:seedance@2.5",
  "positivePrompt": "Create a polished archival photo-restoration reveal for a public-health museum. Full-screen landscape presentation of a genuine-looking 1952 documentary photograph from a rural British maternity clinic: a district midwife in a practical uniform stands beside a young mother holding a swaddled newborn, with a painted iron cot, enamel washbasin and frosted window in the modest clinic room. Begin with the photograph severely but realistically damaged: monochrome silver-gelatin print, faded contrast, warm paper discoloration, branching creases, dust, emulsion scratches, one torn upper corner and several small areas of image loss. Keep the composition and every person completely still throughout; this is restoration of one historical photograph, not animation of the photographed scene. Over five seconds, a precise vertical restoration pass travels smoothly from left to right. Behind the moving boundary, repair tears and scratches, recover missing photographic detail, balance exposure and gently colorize with historically plausible restrained tones: pale sage walls, cream enamelware, muted navy midwife uniform, natural skin tones, soft white cotton and subdued brown furniture. Preserve authentic 1950s film grain, lens softness, facial identity, period details and documentary realism; do not over-sharpen, beautify faces, modernize objects or make colors glossy. The boundary itself should be subtle, visible only through the contrast between damaged and restored areas. Locked camera, no zoom, no added typography, no frames, no interface, no split-screen labels. End on the fully repaired, colorized photograph held cleanly for the final second. Synchronized audio: quiet museum-room ambience, delicate paper crackle at the beginning, then a soft dry brush-like sweep following the restoration boundary, with the crackle fading into silence as the image becomes complete.",
  "width": 1280,
  "height": 720,
  "duration": 5,
  "settings": {
    "audio": true
  }
}
```

---

### first-frame (first-frame)

[Watch video](https://assets.runware.ai/examples/bytedance-seedance-2-5/dfa12fe3-3e07-461c-ab8b-0cf5cb5b9af9.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "2c51550a-8175-47ea-a1cf-dd5463888a0e",
  "model": "bytedance:seedance@2.5",
  "positivePrompt": "Create a single continuous six-second luxury resort cleanup shot beginning from the exact composition and moment shown in @Image1. Preserve the geothermal pool, architecture, volcanic-stone wall, paving geometry, dawn lighting, color, reflections, and restrained photorealistic style. The camera begins a very slow, stable dolly to the right with subtle natural parallax. Steam rises from the turquoise water and drifts laterally across the terrace, briefly covering the cleaning cart and coiled yellow hose as a fully motivated soft wipe. While they are concealed, remove both maintenance items completely. As the steam thins, reveal a seamless clean terrace: generatively reconstruct the missing pale limestone slabs, grout lines, damp surface texture, reflections, contact shadows, and pool-edge details so they remain spatially consistent during the camera move. The equipment must not reappear, leave fragments, distort the paving, or create a visible transition boundary. Water ripples gently, steam curls naturally, and distant grasses move slightly in the breeze. No people, no text, no logos, no cut, no speed ramp, no magical disappearance effect, and no changes to the permanent environment. Polished widescreen hospitality campaign footage with calm, cool, restorative mood.",
  "resolution": "1080p",
  "duration": 6,
  "inputs": {
    "frameImages": [
      "https://assets.runware.ai/assets/inputs/80dbfe8c-38c7-47cb-bc6d-d77aa2f5a878.jpg"
    ]
  }
}
```

---

### first-last-frame (first-last-frame)

[Watch video](https://assets.runware.ai/examples/bytedance-seedance-2-5/8c1fe97b-9d39-409d-bc59-c9bcdddba9f5.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "886aab04-306e-4d23-bfc6-81d1b85bf485",
  "model": "bytedance:seedance@2.5",
  "positivePrompt": "Create a polished 10-second horizontal product colorway reveal using @Image1 as the exact first frame and @Image2 as the exact last frame. This is a commercial launch asset for a portable four-track cassette recorder, not a redesign. Keep the camera locked in the same three-quarter overhead composition and preserve the recorder's silhouette, proportions, cassette window, cassette, knobs, faders, transport icons, hardware, tabletop, lighting direction and shadows throughout.\n\nBegin with a brief confident hold on the warm ivory recorder from @Image1. The cassette reels start turning slowly and naturally. A narrow band of saturated amber sunlight then travels smoothly from the left edge of the recorder to the right, like light passing through a moving warm color filter. Directly behind the advancing light boundary, only the painted outer housing changes from warm ivory enamel to the rich matte persimmon-orange colorway shown in @Image2. Make the transition clean, continuous and materially convincing: enamel pigment changes while highlights, fine surface texture, seams and reflections remain stable. The smoked cassette window, clear cassette, dark brown knobs, ivory faders, silver hardware and white markings must never change color or shape. No liquid paint, splashes, smoke, particles, disassembly or fantasy morphing.\n\nAs the color boundary completes its pass, let the amber light soften back into the original late-afternoon illumination. End on a steady hero hold that matches @Image2 exactly, with the new persimmon housing fully resolved and the cassette reels easing to a stop. Premium analog-audio product cinematography, tactile cork and enamel textures, warm optimistic mood, precise continuity, subtle synchronized reel whirr and one quiet mechanical click, no speech, no music, no added titles, no logo and no camera shake.",
  "resolution": "720p",
  "duration": 10,
  "inputs": {
    "frameImages": [
      {
        "image": "https://assets.runware.ai/assets/inputs/091cd9ef-f1c8-4205-9404-48df2919ba1e.jpg",
        "frame": "first"
      },
      {
        "image": "https://assets.runware.ai/assets/inputs/6ef2f6f5-2314-469f-8729-360b42d11187.jpg",
        "frame": "last"
      }
    ]
  }
}
```

---

### reference-to-video (reference-to-video)

[Watch video](https://assets.runware.ai/examples/bytedance-seedance-2-5/affbed3b-93d4-4fd3-9a61-f8bc01de18f5.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "80b143bf-63dd-404a-9c84-4b3a95a9478c",
  "model": "bytedance:seedance@2.5",
  "positivePrompt": "Create a newly generated, photorealistic documentary video that performs a convincing scene replacement for the beekeeper in @Image1. Preserve her recognizable face, age, body proportions, white protective suit, lowered mesh veil, tan gloves, boots and exact wooden honeycomb frame throughout. Place her naturally inside the dawn rooftop apiary shown in @Video1, replacing the plain studio background with that location. Match @Video1's slow stabilized left-to-right dolly movement, waist-height camera, lens perspective, shallow natural depth of field, apricot sunrise, cool blue city haze, hive layout and restrained observational art direction. She stands between two hive rows, looks down to inspect the honeycomb frame, subtly tilts it toward the sunrise and then raises her gaze toward the nearest hive. Keep her motion calm, skilled and physically plausible. Add accurate contact shadows, warm rim light on the suit, faint reflections on the veil mesh and coherent occlusion as a few bees pass in front of and behind her. Preserve realistic rooftop scale; do not create oversized insects or a fantasy environment. Synchronize the scene with @Audio1, using its bee hum, breeze, distant traffic and bird ambience as the finished soundtrack. Continuous single shot, no cuts, no dialogue, no captions, no logos, no graphic overlays.",
  "resolution": "720p",
  "duration": 8
}
```

---

### video-edit (video-edit)

[Watch video](https://assets.runware.ai/examples/bytedance-seedance-2-5/040eb9e6-5c30-47c3-9303-3093b40f8d5b.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "6d45a666-6d0c-4a33-9726-2acded0112b3",
  "model": "bytedance:seedance@2.5",
  "positivePrompt": "Edit the existing source clip @Video1 for its full original duration; do not extend it or create any footage beyond its ending. Remove the turquoise GPS tracking collar, buckle, electronics housing, and antenna from the Iberian lynx in every frame. Reconstruct the concealed neck and shoulder area with anatomically accurate tawny fur, natural spots, guard hairs, and soft undercoat. Match the direction, compression, lift, and subtle secondary motion of the surrounding fur as the lynx walks and turns its head. Maintain clean temporal consistency through changing angles, motion blur, shifting sunlight, and foreground grasses that briefly occlude the neck. Preserve the lynx's exact anatomy, gait, expression, markings, position, and movement. Preserve the original camera tracking, handheld character, framing, depth of field, vegetation, lighting, color, duration, pacing, and natural location audio. Make no other visible changes. The finished result should look like authentic unaltered natural-history documentary footage with no collar, no antenna, no seams, no flicker, and no cleanup artifacts.",
  "resolution": "720p",
  "duration": "auto",
  "settings": {
    "operation": "edit"
  },
  "inputs": {
    "video": "https://assets.runware.ai/assets/inputs/eabcbc95-4511-488f-9ec4-2bd82cb02c54.mp4"
  }
}
```

---

### video-extend (video-extend)

[Watch video](https://assets.runware.ai/examples/bytedance-seedance-2-5/b6269cb9-715f-42df-be65-37c6e9ba8340.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "494c17c1-97a6-4ff4-92e8-1dbda451f2c0",
  "model": "bytedance:seedance@2.5",
  "positivePrompt": "Continue @Video1 immediately after its final frame without modifying the source footage. The first generated frame must match the final frame exactly, preserving the aerial composition, quarry geometry, crop and slow forward push. Over the continuation, progressively restyle the photograph into a sophisticated hand-carved reduction linocut illustration for a landscape architecture awards film. Begin with subtle carved hatching appearing in the pale limestone at the lower edge, then let the printmaking treatment travel naturally across the terraces, pond, road, trees and sky while every feature remains fixed in place. Replace photographic textures with tactile ink shapes, visible gouge marks, irregular registration and layered relief-print grain. Use a restrained mineral palette of limestone cream, carbon black, oxidized clay red and muted lichen green; render the pond as dense black ink with thin cream reflections. Keep the original tonal hierarchy and recognizable rehabilitation design. Maintain the same smooth forward motion throughout, with no cut, reframing, new objects or spatial warping. By the final moment, the entire landscape is a fully resolved linocut print with crisp carved detail and subtle handmade ink variation, held in the same composition. No lettering, logos, borders or mockup presentation.",
  "resolution": "1080p",
  "duration": 6,
  "settings": {
    "operation": "extend"
  },
  "inputs": {
    "video": "https://assets.runware.ai/assets/inputs/d8022cd8-1920-4457-badd-713e5a01c0ec.mp4"
  }
}
```