---
title: SkyReels V4 | Runware Docs
url: https://runware.ai/docs/models/skywork-skyreels-v4
description: Unified multimodal video model for generation, inpainting, and editing with synchronized audio
---
# SkyReels V4

![SkyReels V4](https://assets.runware.ai/covers/skywork-skyreels-v4.jpg)

SkyReels V4 is a unified multimodal video foundation model for joint video-audio generation, inpainting, and editing. It accepts text, images, video clips, masks, and audio references, and supports cinematic outputs up to 1080p, 32 FPS, and 15 seconds with synchronized audio, making it suitable for prompt-driven generation as well as guided editing workflows.

- **Status**: live
- **ID**: `skywork:skyreels@v4`
- Try in Playground
- **Creator**: Skywork
- **Release Date**: April 24, 2026
- **Capabilities**: Text to Video, Image to Video, Video to Video, Audio to Video, Edit, Extend, Checkpoint, Audio

## Pricing

$0.11 from, per second · ~4m 3s per request

**Rates** (per second)

*Text/image to video*

- 480p: `$0.11`
- 720p: `$0.14`
- 1080p: `$0.35`

*Video to video*

- 480p: `$0.18`
- 720p: `$0.25`
- 1080p: `$0.625`

**7 example generations**

- [first-frame](https://runware.ai/docs/models/skywork-skyreels-v4/examples#example-colorized-lighthouse-heritage-clip) · 720p · 5s · audio: `$0.704` (~2m 26s)
- [multiple-keyframes](https://runware.ai/docs/models/skywork-skyreels-v4/examples#example-volcanologist-keynote-motion-headshot) · 480p · 9s: `$0.994` (~4m 3s)
- [audio-driven](https://runware.ai/docs/models/skywork-skyreels-v4/examples#example-vertical-seed-vault-technician-clip) · 720p · 8s: `$1.12` (~8m 58s)
- [video-extend](https://runware.ai/docs/models/skywork-skyreels-v4/examples#example-antique-auction-video-cleanup) · 720p · 5s: `$1.26` (~9m 28s)
- [first-last-frame](https://runware.ai/docs/models/skywork-skyreels-v4/examples#example-vintage-neon-restoration-social-reel) · 1080p · 6s: `$2.11` (~3m 46s)
- [text-to-video](https://runware.ai/docs/models/skywork-skyreels-v4/examples#example-subsea-cable-documentary-teaser) · 1080p · 6s · audio: `$2.11` (~3m 44s)
- [reference-to-video](https://runware.ai/docs/models/skywork-skyreels-v4/examples#example-court-stenographer-thriller-title-sequence) · 1080p · 8s: `$3.77` (~8m 55s)

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

## Compatibility & Validation

Inside `inputs`, `referenceImages` containing type `grid` cannot be used with `referenceVideos` containing type `reference`.

---

Inside `inputs`, `frameImages` cannot be used with `referenceImages`.

---

Inside `inputs`, `frameImages` cannot be used with `referenceVideos`.

---

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

---

When `inputs.referenceImages` includes type `grid`, `inputs.referenceImages.audio` cannot be used.

---

When `inputs.referenceImages` includes type `grid`, `inputs.referenceImages` is limited to 1 item.

---

When `inputs.referenceImages` includes type `image`, `inputs.referenceImages` is limited to 3 items.

---

When `inputs.referenceVideos` has every item set to type `extend`, `inputs.referenceImages` cannot be used.

---

When `inputs.referenceVideos` includes type `reference`, `settings.audio` cannot be used.

---

When `resolution` is provided, `inputs.frameImages`, `inputs.referenceImages`, or `inputs.referenceVideos` is required.

---

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

---

The following dimension combinations are supported:

| Configuration | Dimensions |
| --- | --- |
| `480p (16:9)` | `848x480` |
| `480p (9:16)` | `480x848` |
| `480p (~4:3)` | `736x560` |
| `480p (~3:4)` | `560x736` |
| `480p (1:1)` | `640x640` |
| `720p (16:9)` | `1280x720` |
| `720p (9:16)` | `720x1280` |
| `720p (~4:3)` | `1088x832` |
| `720p (~3:4)` | `832x1088` |
| `720p (1:1)` | `960x960` |
| `1080p (16:9)` | `1920x1088` |
| `1080p (9:16)` | `1088x1920` |
| `1080p (~4:3)` | `1664x1248` |
| `1080p (~3:4)` | `1248x1664` |
| `1080p (1:1)` | `1440x1440` |

## Notes

SkyReels V4 supports multiple generation modes:

- **Text-to-Video:** Provide a prompt with dimensions (or resolution) and duration.
- **Image-to-Video:** Provide reference images via `inputs.referenceImages` with type `image` (up to 3 groups, 1–5 images each) or `grid` (1 group, 1 collage image).
- **Omni (Reference):** Provide a reference video via `inputs.referenceVideos` with type `reference` for motion guidance, subject replacement, or editing. Input video audio is carried to output.
- **Video Extension:** Provide a reference video with type `extend`. Cannot be combined with reference images. Returns only the generated extension, not the full appended video.

**Reference video requirements:** Input videos must have at least **1,866,240 pixels** (e.g. 1920 × 972) for the selected output resolution. Max 15 seconds.

**Reference images:** All items in `inputs.referenceImages` must share the same `type`. Each group is linked to the prompt via a `tag` starting with `@` (e.g. `@actor1`, `@style`).

**Audio:** Sound effect generation (`settings.audio`) has no effect in Omni mode. Audio from the input video is used instead.

## Request Parameters

**API Options**

Platform-level options for task execution and delivery.

### [taskType](https://runware.ai/docs/models/skywork-skyreels-v4#request-tasktype)

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

Identifier for the type of task being performed.

### [taskUUID](https://runware.ai/docs/models/skywork-skyreels-v4#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/skywork-skyreels-v4#request-outputtype)

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

Video output type.

**Allowed values**: `URL`

### [outputFormat](https://runware.ai/docs/models/skywork-skyreels-v4#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/skywork-skyreels-v4#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/skywork-skyreels-v4#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/skywork-skyreels-v4#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/skywork-skyreels-v4#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/skywork-skyreels-v4#request-safety)

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

Content safety checking configuration for video generation.

#### [checkContent](https://runware.ai/docs/models/skywork-skyreels-v4#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/skywork-skyreels-v4#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/skywork-skyreels-v4#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/skywork-skyreels-v4#request-includecost)

- **Type**: `boolean`

Include task cost in the response.

### [numberResults](https://runware.ai/docs/models/skywork-skyreels-v4#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/skywork-skyreels-v4#request-inputs-referenceimages)

- **Path**: `inputs.referenceImages`
- **Type**: `array of objects (4 properties)`
- **Min items**: `1`
- **Max items**: `3`
- **Area (each)**: `1.87 MP or more`

Reference image groups for subject, scene, style, or collage-based guidance. All items must share the same type.

#### [type](https://runware.ai/docs/models/skywork-skyreels-v4#request-inputs-referenceimages-type)

- **Path**: `inputs.referenceImages.type`
- **Type**: `string`
- **Required**: true

Reference mode for this group.

**Allowed values**:

- `grid` Single collage image.
- `image` Subject, scene, or style reference.

#### [audio](https://runware.ai/docs/models/skywork-skyreels-v4#request-inputs-referenceimages-audio)

- **Path**: `inputs.referenceImages.audio`
- **Type**: `string`

Audio input (UUID or URL). Voice timbre reference for lip-sync. Max 15 seconds.

#### [images](https://runware.ai/docs/models/skywork-skyreels-v4#request-inputs-referenceimages-images)

- **Path**: `inputs.referenceImages.images`
- **Type**: `array of strings`
- **Required**: true
- **Min items**: `1`
- **Max items**: `5`

Image inputs (UUID, URL, Data URI, or Base64) for this reference group.

#### [tag](https://runware.ai/docs/models/skywork-skyreels-v4#request-inputs-referenceimages-tag)

- **Path**: `inputs.referenceImages.tag`
- **Type**: `string`
- **Required**: true

Group identifier that links this reference to a mention in the prompt. Must start with @ (e.g. @actor1, @style).

### [frameImages](https://runware.ai/docs/models/skywork-skyreels-v4#request-inputs-frameimages)

- **Path**: `inputs.frameImages`
- **Type**: `array of strings or objects`
- **Min items**: `1`
- **Max items**: `8`

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.
- **3+ images**: First and last frames, with intermediate images evenly spaced between.

**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"
  }
]
```

**Multiple frames:** With three or more images, the first and last are anchored while intermediate frames are evenly distributed across the timeline.

```json
"frameImages": [
  {
    "image": "aac49721-1964-481a-ae78-8a4e29b91402",
    "frame": "first"
  },
  "c00abf5f-6cdb-4642-a01d-1bfff7bc3cf7",
  {
    "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/skywork-skyreels-v4#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/skywork-skyreels-v4#request-inputs-frameimages-format-2-frame)

- **Path**: `inputs.frameImages.frame`
- **Type**: `string | integer`
- **Min**: `-1`

Target frame position for the image. Accepts a named position like `first` or `last`, or a zero-based frame index (-1 for the last frame).

**Allowed values**: `first` `last`

### [referenceVideos](https://runware.ai/docs/models/skywork-skyreels-v4#request-inputs-referencevideos)

- **Path**: `inputs.referenceVideos`
- **Type**: `array of objects (3 properties)`
- **Min items**: `items: 1`
- **Area (each)**: `1.87 MP or more`

Reference video for motion guidance, subject replacement, editing, or video extension.

#### [type](https://runware.ai/docs/models/skywork-skyreels-v4#request-inputs-referencevideos-type)

- **Path**: `inputs.referenceVideos.type`
- **Type**: `string`
- **Required**: true

Reference mode for the video.

**Allowed values**:

- `reference` Motion reference, subject replacement, or editing. Input video audio is carried to output.
- `extend` Video extension.

#### [tag](https://runware.ai/docs/models/skywork-skyreels-v4#request-inputs-referencevideos-tag)

- **Path**: `inputs.referenceVideos.tag`
- **Type**: `string`
- **Required**: true

Video reference identifier that links to a mention in the prompt. Must start with @ (e.g. @video1).

#### [video](https://runware.ai/docs/models/skywork-skyreels-v4#request-inputs-referencevideos-video)

- **Path**: `inputs.referenceVideos.video`
- **Type**: `string`
- **Required**: true

Video input (UUID or URL). Max 15 seconds. Minimum 1,866,240 total pixels (e.g. 1920×972).

**Core Parameters**

Primary parameters that define the task output.

### [model](https://runware.ai/docs/models/skywork-skyreels-v4#request-model)

- **Type**: `string`
- **Required**: true
- **Value**: `skywork:skyreels@v4`

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/skywork-skyreels-v4#request-positiveprompt)

- **Type**: `string`
- **Required**: true
- **Max**: `5120`

Text prompt describing elements to include in the generated output.

**Learn more** (1 resource):

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

### [width](https://runware.ai/docs/models/skywork-skyreels-v4#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/skywork-skyreels-v4#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/skywork-skyreels-v4#request-resolution)

- **Type**: `string`
- **Default**: `1080p`

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

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

### [duration](https://runware.ai/docs/models/skywork-skyreels-v4#request-duration)

- **Type**: `integer`
- **Min**: `3`
- **Max**: `15`
- **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).

**Settings**

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

### [audio](https://runware.ai/docs/models/skywork-skyreels-v4#request-settings-audio)

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

Sound effect generation for the output video.

### [promptExtend](https://runware.ai/docs/models/skywork-skyreels-v4#request-settings-promptextend)

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

Automatic prompt expansion and refinement via LLM for higher visual fidelity and alignment.

## Response Parameters

### [taskType](https://runware.ai/docs/models/skywork-skyreels-v4#response-tasktype)

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

Identifier for the type of task this response belongs to.

### [taskUUID](https://runware.ai/docs/models/skywork-skyreels-v4#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/skywork-skyreels-v4#response-videouuid)

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

UUID of the output video.

### [videoURL](https://runware.ai/docs/models/skywork-skyreels-v4#response-videourl)

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

URL of the output video.

### [videoBase64Data](https://runware.ai/docs/models/skywork-skyreels-v4#response-videobase64data)

- **Type**: `string`

Base64-encoded video data.

### [videoDataURI](https://runware.ai/docs/models/skywork-skyreels-v4#response-videodatauri)

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

Data URI of the output video.

### [NSFWContent](https://runware.ai/docs/models/skywork-skyreels-v4#response-nsfwcontent)

- **Type**: `boolean`

Flag indicating if NSFW content was detected.

### [cost](https://runware.ai/docs/models/skywork-skyreels-v4#response-cost)

- **Type**: `float`

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

## Examples

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

[Watch video](https://assets.runware.ai/examples/skywork-skyreels-v4/bf9c3a39-291d-4b2a-be92-34646f195725.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "17ac60fc-d552-4dd6-b934-70c6669aca62",
  "model": "skywork:skyreels@v4",
  "positivePrompt": "Create a polished six-second cinematic teaser for a prestige technology documentary titled “BENEATH THE SIGNAL.” One continuous, constantly developing macro shot inside a vast, dark submarine-cable landing station. Begin in near-blackness on an extreme close-up of hair-thin glass fibers carrying rapid pulses of electric cobalt light. The camera races backward along the glowing fibers as they twist into a thick armored communications cable; copper shielding, black polymer, salt-stained steel, warning-orange clamps, condensation, and drifting dust snap into view with crisp photoreal detail. Follow the cable as it whips through heavy industrial rollers, then sweep low across the concrete floor while synchronized blue data pulses race toward a monumental cable drum. In the final two seconds, the moving light traces converge across the dark face of the drum and resolve into large, sharply legible condensed uppercase lettering: “BENEATH THE SIGNAL.” Hold the completed title just long enough to read as a final low-frequency pulse travels through it. Brutalist industrial art direction, deep blacks, electric cobalt, oxidized copper, and restrained safety orange; hard side lighting, volumetric haze, shallow depth of field shifting into a crisp final title, subtle anamorphic flare, premium documentary finish, controlled motion blur. Compose for widescreen landscape with the title centered inside safe margins. No people, logos, subtitles, credits, interface graphics, or additional words. Synchronized sound design: rising electrical hum, delicate fiber-optic ticks, roller clanks, tightening cable strain, a fast data-like shimmer, and one deep resonant metallic impact on the final title reveal.",
  "width": 1920,
  "height": 1088,
  "duration": 6,
  "settings": {
    "audio": true
  }
}
```

---

### first-frame (first-frame)

[Watch video](https://assets.runware.ai/examples/skywork-skyreels-v4/7bea0630-61b5-434a-8c8c-86dfae044afb.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "6e50fa8e-38d6-401b-b5d5-b149bce3bca9",
  "model": "skywork:skyreels@v4",
  "positivePrompt": "Continue naturally from the supplied first frame as a restrained living-history moment inside the 1937 lighthouse lantern room. Preserve the two women's identities, period clothing, restored photographic detail, room geometry, Fresnel lens construction, and muted historically researched color palette. The keeper slowly turns the brass crank by a few inches; the enormous glass lens begins a gentle, mechanically plausible rotation. Her daughter completes one careful wiping motion across a prism and briefly glances toward the moving lens. Morning light refracts through the rotating glass, sending a soft band of warm illumination across their faces and the painted wall. Beyond the windows, the distant sea moves subtly and one curtain edge stirs in a light draft. Use a very slow, steady documentary push-in that emphasizes the foreground lens, layered depth, and intimate conservation detail. Natural human timing, minimal motion, stable faces and hands, authentic 1930s atmosphere, no cuts, no added people, no modern objects, no lettering, no dramatic fantasy effects.",
  "resolution": "720p",
  "duration": 5,
  "settings": {
    "audio": true
  },
  "inputs": {
    "frameImages": [
      "https://assets.runware.ai/assets/inputs/6265784c-73b5-4d16-991f-374eda97a98e.jpg"
    ]
  }
}
```

---

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

[Watch video](https://assets.runware.ai/examples/skywork-skyreels-v4/bade6d5c-ce1c-42f8-99d1-d207249cbf5e.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "b6a9c96d-1fa4-4231-a17d-134b2d0bedd7",
  "model": "skywork:skyreels@v4",
  "positivePrompt": "Create a polished vertical short-form social media ad for a specialist vintage neon restoration studio, using the supplied first and last frames as exact visual anchors. Begin on the weathered, unlit abstract sunburst arrow sign and the restorer wiping its chipped enamel at the workbench. In one fluid accelerated transition, the camera slowly pushes forward and lowers while her hands work with confident precision: grime clears beneath the cloth, loose wiring is secured, and reflections glide over the repaired glass tubes. She reaches to the transformer switch; a crisp electrical click triggers the neon to flicker alive section by section—cyan first, then coral, then warm amber—until the restored sign glows steadily. Resolve naturally into the supplied final frame with the same woman, sign, clothing, tools, and workshop layout. Premium tactile documentary advertising style with rhythmic process-film energy, realistic hand movement, physically accurate glass and electrical illumination, rich colored reflections, deep industrial shadows, subtle floating dust, and smooth camera motion. Compose every moment for a 1080p vertical social feed with the sign and craftsperson remaining clear in the central safe area. Synchronized sound design: cloth brushing painted metal, small tool clinks, transformer hum, switch click, brief neon buzz, then a warm steady electrical ambience. No dialogue, captions, lettering, logos, borders, interface elements, or watermark.",
  "resolution": "1080p",
  "duration": 6,
  "inputs": {
    "frameImages": [
      {
        "image": "https://assets.runware.ai/assets/inputs/4ef4929f-bf31-4d0c-ac05-36182657edf6.jpg",
        "frame": "first"
      },
      {
        "image": "https://assets.runware.ai/assets/inputs/39e4c885-5626-4c06-81b0-c727c424ad1b.jpg",
        "frame": "last"
      }
    ]
  }
}
```

---

### multiple-keyframes (multiple-keyframes)

[Watch video](https://assets.runware.ai/examples/skywork-skyreels-v4/ee97aa04-1b98-49e4-92bb-71875b943723.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "549b602d-41d7-4d0b-92d8-db88db1d8220",
  "model": "skywork:skyreels@v4",
  "positivePrompt": "Create a premium retouched motion headshot for a scientific conference keynote profile, using the three supplied portraits as keyframes of one continuous nine-second action. Begin exactly from the first frame: the volcanologist holds a gentle three-quarter pose with her chin lowered and gaze down-left. She breathes naturally, then slowly raises her chin and turns toward camera. Arrive naturally at @image1 midway through the movement, with her eyes meeting the lens and her expression still attentive and neutral. Continue without a cut as her shoulders settle, her gaze becomes assured and a subtle closed-mouth smile develops, ending exactly on the supplied final frame. Preserve the woman’s identity, age, facial proportions, hairstyle, wardrobe, earrings, background, camera position and lighting throughout. Apply stable, high-end editorial headshot enhancement across the full clip: balance complexion and temporary redness, gently soften under-eye shadows, control facial shine, remove isolated lint and distracting flyaway hairs, and maintain clean catchlights. Preserve pores, fine lines, natural asymmetry and realistic mature skin; no plastic smoothing, glamor makeup, face reshaping or age reduction. Keep movement minimal and credible, with tiny blinks, restrained breathing and physically consistent hair. Locked-off camera, chest-up landscape composition, shallow but stable depth of field, soft cool-gray fill with a restrained copper rim light, charcoal basalt-gray palette, dignified and quietly optimistic mood. No text, graphics, cuts, zooms, flicker, identity drift or wardrobe changes.",
  "resolution": "480p",
  "duration": 9,
  "inputs": {
    "frameImages": [
      {
        "image": "https://assets.runware.ai/assets/inputs/fbf12db9-c33d-4b5e-b946-52e2cd8910bb.jpg",
        "frame": "first"
      },
      {
        "image": "https://assets.runware.ai/assets/inputs/88139be7-d582-499a-9eb7-2be5206e253c.jpg",
        "frame": 108
      },
      {
        "image": "https://assets.runware.ai/assets/inputs/2a3d5514-3363-420f-8127-5188f7ba17b4.jpg",
        "frame": "last"
      }
    ]
  }
}
```

---

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

[Watch video](https://assets.runware.ai/examples/skywork-skyreels-v4/f16e3f56-7010-49e4-ab47-bc19662d103b.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "9cfada7f-7ed0-4940-add5-897897f03101",
  "model": "skywork:skyreels@v4",
  "positivePrompt": "Create a finished eight-second cinematic title sequence for a prestige legal conspiracy thriller titled “THE RECORD.” Use @actor1 as the court stenographer and preserve her facial identity, age, hair and dark-red glasses. Replace the performer in @video1 with @actor1; @actor1 must follow the exact hand movements, seated posture, eye line, timing and restrained performance from @video1. Preserve the driving clip’s deliberate camera push and synchronized original production audio.\n\nSet the action in a deserted municipal courtroom after midnight. Begin with tactile macro detail of @actor1’s fingers hammering the cream and black stenotype keys, transcript paper trembling beneath them. The camera advances toward her face while institutional fluorescent fixtures extinguish sequentially behind her, leaving sickly green light across her skin and deep oxblood shadows in the empty gallery. Fine dust hangs in the light. She stops mid-word, slowly raises her eyes toward the sealed double doors, then tears the transcript strip free with sudden controlled force.\n\nOn the final beat, let the torn bone-white paper whip across the lens as a practical wipe into near-black. The paper settles vertically in the center of frame and the typed uppercase title “THE RECORD” becomes visible in crisp black monospaced letters, with one thin red evidence line beneath it. Hold the title long enough to read. Minimal, severe typography only; no logos, credits, subtitles or additional copy. Photorealistic cinematic texture, restrained 35mm grain, shallow depth of field, precise highlights on machinery, institutional green, bone paper, oxblood and black palette, tense procedural mood, polished premium-streaming finish.",
  "resolution": "1080p",
  "duration": 8,
  "inputs": {
    "referenceImages": [
      {
        "type": "image",
        "tag": "@actor1",
        "images": [
          "https://assets.runware.ai/assets/inputs/70648e1c-eff4-47bf-996e-ec645437bff6.jpg"
        ]
      }
    ],
    "referenceVideos": [
      {
        "type": "reference",
        "tag": "@video1",
        "video": "https://assets.runware.ai/assets/inputs/15dc1d11-c508-4ba7-94c2-cc8e457f0129.mp4"
      }
    ]
  }
}
```

---

### audio-driven (audio-driven)

[Watch video](https://assets.runware.ai/examples/skywork-skyreels-v4/8ff29caf-6e81-4d2d-99ed-7ae8913b7f64.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "39884855-18f0-4236-836c-3a4403005168",
  "model": "skywork:skyreels@v4",
  "positivePrompt": "Create a single continuous 8-second vertical documentary b-roll shot featuring @speaker1 inside a working agricultural seed vault. Preserve her identity, clothing, and natural facial features from the reference image. Begin in a waist-up portrait composition as she carefully slides a sealed seed packet into a frosted metal storage drawer. The camera makes a subtle, stable push-in. She looks up toward an off-camera interviewer and delivers the supplied @speaker1 voice line with precise, natural lip-sync, believable jaw movement, restrained hand gestures, and a calm expert demeanor. After speaking, she briefly glances back toward the drawer. Cool cyan freezer light, soft white overhead illumination, faint condensation in the cold air, navy and stainless-steel palette with orange gloves as the only warm accent. Authentic climate-science workplace footage, observational editorial realism, crisp face and eyes, shallow but practical depth of field, gentle handheld micro-movement, no dramatic effects, no cuts, no other people, no captions, no logos, no readable text.",
  "width": 720,
  "height": 1280,
  "duration": 8
}
```

---

### video-extend (video-extend)

[Watch video](https://assets.runware.ai/examples/skywork-skyreels-v4/053902ac-72e4-4c7d-be64-4b46fbef807e.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "8c100ed5-7d7d-415d-9458-11e74cf97547",
  "model": "skywork:skyreels@v4",
  "positivePrompt": "Use @video1 as the source shot. Cleanly remove the fluorescent green temporary inventory card from every visible frame and generatively reconstruct the dark mirrored pedestal beneath it, including physically accurate reflections, edge highlights, fine dust texture, and changing parallax during the camera move. Preserve the antique astronomical clock exactly: its proportions, engraved brass surfaces, moving escapement, hands, patina, and warm gallery lighting must remain unchanged. Continue the action beyond the end of @video1 for five seconds: maintain the same slow, stable rightward camera arc as the clock mechanism keeps ticking, reveal slightly more of the clock's right profile, then ease toward a composed three-quarter catalog view. Match the original lens, shallow depth of field, exposure, midnight-blue background, shadow direction, motion cadence, and uninterrupted ticking ambience. Seamless professional auction-film cleanup with no visible patching, warping, flicker, added objects, labels, or text.",
  "resolution": "720p",
  "duration": 5,
  "inputs": {
    "referenceVideos": [
      {
        "type": "extend",
        "tag": "@video1",
        "video": "https://assets.runware.ai/assets/inputs/515d5984-f917-4ead-894f-f212797a99a9.mp4"
      }
    ]
  }
}
```