---
title: Ray3.2 | Runware Docs
url: https://runware.ai/docs/models/luma-ray3-2
description: Cinematic video model for generation, transformation, motion transfer, and frame-level direction
---
# Ray3.2

![Ray3.2](https://assets.runware.ai/covers/luma-ray3-2.jpg)

Ray3.2 is Luma's flagship video model for turning creative direction into controllable production workflows. It supports text-to-video, image-to-video, and video-to-video generation, with stronger continuity, motion transfer, camera motion transfer, character transformation, relighting, environment change, and product-swap workflows. It is built for cinematic-quality output, multi-keyframe control inside a single clip, and Modify Video V2 workflows that preserve performance, lighting, and scene structure while transforming existing footage.

- **Status**: live
- **ID**: `luma:ray@3.2`
- Try in Playground
- **Creator**: Luma
- **Release Date**: June 9, 2026
- **Capabilities**: Text to Video, Image to Video, Video to Video, Edit, Checkpoint

## Pricing

$0.012 from, per second · ~1m 25s per request

**Rates**

*Text/image to video*

- 360p · per second: `$0.012`
- 540p · per second: `$0.03`
- 720p · per second: `$0.06`
- 1080p · per second: `$0.24`

*HDR*

- 720p · per second: `$0.12`
- 1080p · per second: `$0.48`

*HDR+EXR*

- 720p · per second: `$0.18`
- 1080p · per second: `$0.72`

*Video edit*

- 360p · per second: `$0.108`
- 540p · per second: `$0.144`
- 720p · per second: `$0.216`
- 1080p · per second: `$0.432`

*Extend*

- 540p · per run: `$0.15`
- 720p · per run: `$0.30`
- 1080p · per run: `$1.20`

*Reframe*

- 320p · per second: `$0.03`
- 540p · per second: `$0.06`
- 720p · per second: `$0.12`
- 1080p · per second: `$0.36`

*Video edit HDR*

- 720p · per second: `$0.432`
- 1080p · per second: `$0.864`

*Video edit HDR+EXR*

- 720p · per second: `$0.648`
- 1080p · per second: `$1.30`

**6 example generations**

- [first-last-frame](https://runware.ai/docs/models/luma-ray3-2/examples#example-geothermal-network-logo-sting) · 540p · 5s: `$0.15` (~37s)
- [multiple-keyframes](https://runware.ai/docs/models/luma-ray3-2/examples#example-artisan-paper-marbling-b-roll) · 720p · 5s: `$0.30` (~1m 2s)
- [video-reframe](https://runware.ai/docs/models/luma-ray3-2/examples#example-planetarium-archive-frame-restoration) · 720p: `$0.725` (~1m 23s)
- [text-to-video](https://runware.ai/docs/models/luma-ray3-2/examples#example-colorized-apiary-newsreel-restoration) · 720p · 5s: `$0.90` (~1m 27s)
- [video-edit](https://runware.ai/docs/models/luma-ray3-2/examples#example-veterinary-ultrasound-field-demo) · 720p: `$1.08` (~3m 14s)
- [first-frame](https://runware.ai/docs/models/luma-ray3-2/examples#example-chamber-orchestra-group-portrait-clip) · 1080p · 5s: `$1.20` (~8m 29s)

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 `frameImages`.

---

When `inputs.frameImages` is provided, `duration` must be `5`.

---

When `inputs.video` is provided, `settings.loop` cannot be used.

---

When `inputs.video` is provided, `duration` cannot be used.

---

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

---

When `resolution` is provided, `inputs.frameImages` or `inputs.video` is required.

---

Inside `settings`, `loop` cannot be used with `hdr`.

---

When `settings.edit` is provided, `width/height` cannot be used.

---

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

---

When `settings.edit.autoControls` is `true`, `settings.edit.strength` and `settings.edit.controls` cannot be used.

---

When `settings.exrExport` is `true`, `settings.hdr` must be `true`.

---

When `settings.hdr` is `true`, dimensions `360×640`, `360×480`, `480×480`, `480×360`, `640×360`, `840×360`, `540×960`, `540×720`, `720×720`, `720×540`, `960×540`, `1260×540` cannot be used.

---

When `settings.hdr` is `true`, `resolution` is limited to `720p`, `1080p`.

---

When `settings.hdr` is `true`, `duration` is limited to `5`.

---

When `settings.loop` is `true`, `duration` is limited to `5`.

---

When `settings.sourcePosition` is provided, `inputs.video`, `width`, and `height` are required.

---

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

---

The following dimension combinations are supported:

| Configuration | Dimensions |
| --- | --- |
| `360p (9:16)` | `360x640` |
| `360p (3:4)` | `360x480` |
| `360p (1:1)` | `480x480` |
| `360p (4:3)` | `480x360` |
| `360p (16:9)` | `640x360` |
| `360p (21:9)` | `840x360` |
| `540p (9:16)` | `540x960` |
| `540p (3:4)` | `540x720` |
| `540p (1:1)` | `720x720` |
| `540p (4:3)` | `720x540` |
| `540p (16:9)` | `960x540` |
| `540p (21:9)` | `1260x540` |
| `720p (9:16)` | `720x1280` |
| `720p (3:4)` | `720x960` |
| `720p (1:1)` | `960x960` |
| `720p (4:3)` | `960x720` |
| `720p (16:9)` | `1280x720` |
| `720p (21:9)` | `1680x720` |
| `1080p (9:16)` | `1080x1920` |
| `1080p (3:4)` | `1080x1440` |
| `1080p (1:1)` | `1440x1440` |
| `1080p (4:3)` | `1440x1080` |
| `1080p (16:9)` | `1920x1080` |
| `1080p (21:9)` | `2520x1080` |

## Request Parameters

**API Options**

Platform-level options for task execution and delivery.

### [taskType](https://runware.ai/docs/models/luma-ray3-2#request-tasktype)

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

Identifier for the type of task being performed.

### [taskUUID](https://runware.ai/docs/models/luma-ray3-2#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/luma-ray3-2#request-outputtype)

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

Video output type.

**Allowed values**: `URL`

### [outputFormat](https://runware.ai/docs/models/luma-ray3-2#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/luma-ray3-2#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/luma-ray3-2#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/luma-ray3-2#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/luma-ray3-2#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/luma-ray3-2#request-safety)

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

Content safety checking configuration for video generation.

#### [checkContent](https://runware.ai/docs/models/luma-ray3-2#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/luma-ray3-2#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/luma-ray3-2#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/luma-ray3-2#request-includecost)

- **Type**: `boolean`

Include task cost in the response.

### [numberResults](https://runware.ai/docs/models/luma-ray3-2#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.

### [frameImages](https://runware.ai/docs/models/luma-ray3-2#request-inputs-frameimages)

- **Path**: `inputs.frameImages`
- **Type**: `array of strings or objects`
- **Min items**: `1`
- **Max items**: `64`

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/luma-ray3-2#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/luma-ray3-2#request-inputs-frameimages-format-2-frame)

- **Path**: `inputs.frameImages.frame`
- **Type**: `string | integer`
- **Min**: `-1`
- **Max**: `120`

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`

### [video](https://runware.ai/docs/models/luma-ray3-2#request-inputs-video)

- **Path**: `inputs.video`
- **Type**: `string`

Source video to edit or reframe (UUID or URL). Editing preserves the source aspect ratio; reframing places it on a new canvas.

**Core Parameters**

Primary parameters that define the task output.

### [model](https://runware.ai/docs/models/luma-ray3-2#request-model)

- **Type**: `string`
- **Required**: true
- **Value**: `luma:ray@3.2`

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/luma-ray3-2#request-positiveprompt)

- **Type**: `string`
- **Required**: true
- **Min**: `1`
- **Max**: `6000`

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/luma-ray3-2#request-width)

- **Type**: `integer`
- **Required**: conditional — Required when settings.sourcePosition is provided
- **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/luma-ray3-2#request-height)

- **Type**: `integer`
- **Required**: conditional — Required when settings.sourcePosition is provided
- **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/luma-ray3-2#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**: `360p` `540p` `720p` `1080p`

### [duration](https://runware.ai/docs/models/luma-ray3-2#request-duration)

- **Type**: `integer`
- **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**: `5` `10`

**Settings**

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

### [edit](https://runware.ai/docs/models/luma-ray3-2#request-settings-edit)

- **Path**: `settings.edit`
- **Type**: `object (8 properties)`

Conditioning controls applied when editing the source video.

#### [strength](https://runware.ai/docs/models/luma-ray3-2#request-settings-edit-strength)

- **Path**: `settings.edit.strength`
- **Type**: `string`

Preservation preset, from strict adherence to free reinterpretation.

**Allowed values**: `adhere_1` `adhere_2` `adhere_3` `flex_1` `flex_2` `flex_3` `reimagine_1` `reimagine_2` `reimagine_3`

#### [autoControls](https://runware.ai/docs/models/luma-ray3-2#request-settings-edit-autocontrols)

- **Path**: `settings.edit.autoControls`
- **Type**: `boolean`

Derive the full conditioning schedule automatically from the source video.

#### [controls](https://runware.ai/docs/models/luma-ray3-2#request-settings-edit-controls)

- **Path**: `settings.edit.controls`
- **Type**: `object (5 properties)`

Per-signal conditioning controls.

##### [depthBlur](https://runware.ai/docs/models/luma-ray3-2#request-settings-edit-controls-depthblur)

- **Path**: `settings.edit.controls.depthBlur`
- **Type**: `float`
- **Min**: `0`
- **Max**: `1`

Scene-geometry depth blur. Higher values allow more freedom.

##### [face](https://runware.ai/docs/models/luma-ray3-2#request-settings-edit-controls-face)

- **Path**: `settings.edit.controls.face`
- **Type**: `boolean`

Enable face-identity conditioning.

##### [normalsAugmentation](https://runware.ai/docs/models/luma-ray3-2#request-settings-edit-controls-normalsaugmentation)

- **Path**: `settings.edit.controls.normalsAugmentation`
- **Type**: `float`
- **Min**: `0`
- **Max**: `1`

Surface-geometry normals augmentation. Higher values allow more reinterpretation.

##### [poseStrength](https://runware.ai/docs/models/luma-ray3-2#request-settings-edit-controls-posestrength)

- **Path**: `settings.edit.controls.poseStrength`
- **Type**: `string`

Skeleton conditioning strength.

**Allowed values**: `precise` `coarse`

##### [trajectorySparsity](https://runware.ai/docs/models/luma-ray3-2#request-settings-edit-controls-trajectorysparsity)

- **Path**: `settings.edit.controls.trajectorySparsity`
- **Type**: `float`
- **Min**: `0`
- **Max**: `1`

Motion-anchoring sparsity. Higher values use fewer anchors.

### [exrExport](https://runware.ai/docs/models/luma-ray3-2#request-settings-exrexport)

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

Also export an OpenEXR frame sequence alongside the MP4.

### [hdr](https://runware.ai/docs/models/luma-ray3-2#request-settings-hdr)

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

Render in high dynamic range.

### [loop](https://runware.ai/docs/models/luma-ray3-2#request-settings-loop)

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

Loop the generated video seamlessly.

### [sourcePosition](https://runware.ai/docs/models/luma-ray3-2#request-settings-sourceposition)

- **Path**: `settings.sourcePosition`
- **Type**: `object (4 properties)`

Placement of the source video within the output canvas, as fractions of the canvas size.

#### [width](https://runware.ai/docs/models/luma-ray3-2#request-settings-sourceposition-width)

- **Path**: `settings.sourcePosition.width`
- **Type**: `float`
- **Required**: true
- **Max**: `2`
- **Paired with**: height

Width of the source video, as a fraction of the canvas width.

#### [height](https://runware.ai/docs/models/luma-ray3-2#request-settings-sourceposition-height)

- **Path**: `settings.sourcePosition.height`
- **Type**: `float`
- **Required**: true
- **Max**: `2`
- **Paired with**: width

Height of the source video, as a fraction of the canvas height.

#### [x](https://runware.ai/docs/models/luma-ray3-2#request-settings-sourceposition-x)

- **Path**: `settings.sourcePosition.x`
- **Type**: `float`
- **Required**: true
- **Min**: `-2`
- **Max**: `2`

Horizontal position of the source video, as a fraction of the canvas width.

#### [y](https://runware.ai/docs/models/luma-ray3-2#request-settings-sourceposition-y)

- **Path**: `settings.sourcePosition.y`
- **Type**: `float`
- **Required**: true
- **Min**: `-2`
- **Max**: `2`

Vertical position of the source video, as a fraction of the canvas height.

## Response Parameters

### [taskType](https://runware.ai/docs/models/luma-ray3-2#response-tasktype)

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

Identifier for the type of task this response belongs to.

### [taskUUID](https://runware.ai/docs/models/luma-ray3-2#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/luma-ray3-2#response-videouuid)

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

UUID of the output video.

### [videoURL](https://runware.ai/docs/models/luma-ray3-2#response-videourl)

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

URL of the output video.

### [videoBase64Data](https://runware.ai/docs/models/luma-ray3-2#response-videobase64data)

- **Type**: `string`

Base64-encoded video data.

### [videoDataURI](https://runware.ai/docs/models/luma-ray3-2#response-videodatauri)

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

Data URI of the output video.

### [NSFWContent](https://runware.ai/docs/models/luma-ray3-2#response-nsfwcontent)

- **Type**: `boolean`

Flag indicating if NSFW content was detected.

### [cost](https://runware.ai/docs/models/luma-ray3-2#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/luma-ray3-2/65622403-aa54-4293-a44e-35870b3fde8c.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "8fe7ec2c-0779-4dca-abad-1c60e14baa22",
  "model": "luma:ray@3.2",
  "positivePrompt": "Create a single continuous shot resembling expertly restored and historically colorized 1947 agricultural training-film footage. Inside a small rural cooperative honey house, a beekeeper in a cream rolled-sleeve work shirt, faded blue trousers, and a wax-marked canvas apron steadily turns the wooden hand crank of a galvanized steel honey extractor. The extractor basket rotates smoothly while thick amber honey glints through a small inspection opening. Stacked wooden hive frames, an enamel pail, muslin cloth, and sage-painted timber walls establish an authentic working apiary with no modern objects. Locked-off tripod camera at chest height, medium-wide landscape composition, natural late-afternoon window light, warm honey amber, muted sage, dusty denim blue, shell cream, and restrained skin tones. The crank completes one exact rotation cycle; the beekeeper's hands, posture, extractor position, background activity, and lighting return precisely to their opening state so the final frame reconnects invisibly to the first. Continuous purposeful motion, no cuts or camera movement. Convincing archival restoration: stable registration, repaired exposure, clean detail, temporally consistent colorization, subtle fine 16mm grain, gentle period-lens softness and halation, preserved natural motion cadence. No scratches, tears, dust bursts, frame jumps, flicker, color breathing, split screen, borders, captions, logos, or text.",
  "width": 1280,
  "height": 720,
  "duration": 5,
  "settings": {
    "loop": true
  }
}
```

---

### first-frame (first-frame)

[Watch video](https://assets.runware.ai/examples/luma-ray3-2/4aa60a28-07d3-4c76-82b0-beff1c22f0a9.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "6a868e8a-5e3f-4945-b325-8fa5441007f5",
  "model": "luma:ray@3.2",
  "positivePrompt": "Continue naturally from the opening portrait setup in one uninterrupted cinematic shot. A charismatic female conductor in a sharply tailored ivory dinner jacket hurries into frame from the right, carrying a slim black baton. She crosses the clear stage space and steps into the open position at center-right among the musicians. The ensemble briefly breaks into warm, genuine smiles at her late arrival; the cellist shifts slightly to make room and the clarinetist turns back toward camera. The conductor straightens her jacket, raises the baton vertically against her chest, and settles into a confident formal portrait pose with the full group. Everyone finishes facing the lens, composed and still, instruments remaining consistent and correctly held. Preserve the historic theater, oxblood velvet curtains, walnut floor, warm amber theatrical light, teal-toned shadows, refined black concert wardrobe, and premium season-campaign photography. Subtle slow camera push-in, natural fabric and hair motion, believable footfalls and body weight, stable faces, polished cinematic realism, no cuts, no lettering, no logos.",
  "resolution": "1080p",
  "duration": 5,
  "inputs": {
    "frameImages": [
      {
        "image": "https://assets.runware.ai/assets/inputs/cfdb48cf-604d-4c26-b235-957ddc0484e3.jpg",
        "frame": "first"
      }
    ]
  }
}
```

---

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

[Watch video](https://assets.runware.ai/examples/luma-ray3-2/71bab428-96e3-49c3-810f-2cf6bdaf31d2.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "46001cd3-5b15-44bc-bdb6-1f879660857a",
  "model": "luma:ray@3.2",
  "positivePrompt": "Create a polished five-second animated logo sting connecting the supplied first and last keyframes for DEEPLOOP, a district geothermal energy network. Preserve the warm limestone-white paper field, graphite lines, oxide-orange accent, flat vector finish, and locked straight-on 16:9 composition throughout. Begin exactly from the opening image: five calm geological contour lines span the frame around the tiny central borehole marker. A precise pulse travels outward from the orange marker, making the contour lines flex in sequence like a clean seismic reading. The center line then descends vertically as a borehole, curves beneath the surface, and rises again to form a continuous geothermal circulation loop. As the loop completes, the surrounding graphite strata slide, fold, and resolve into bold geometric letter strokes. The transformation should feel engineered rather than magical: measured easing, tidy joins, no particles, no liquid splashes, no camera movement. Reveal the exact uppercase wordmark “DEEPLOOP”; its two O letters settle into the linked oxide-orange loop symbol while three fine contour lines align beneath it. Finish exactly on the supplied final keyframe and hold the completed logo cleanly for the last half-second. Keep all lettering stable, sharp, correctly spelled, and free of additional text.",
  "resolution": "540p",
  "duration": 5,
  "inputs": {
    "frameImages": [
      {
        "image": "https://assets.runware.ai/assets/inputs/cb2d35bd-23f2-43ae-a781-582deaf7d5ec.jpg",
        "frame": "first"
      },
      {
        "image": "https://assets.runware.ai/assets/inputs/4851aad2-8803-44bf-b9c5-412a28fd41be.jpg",
        "frame": "last"
      }
    ]
  }
}
```

---

### multiple-keyframes (multiple-keyframes)

[Watch video](https://assets.runware.ai/examples/luma-ray3-2/b9599e80-53f3-4a17-91a5-ad8f489b6769.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "18f4d187-a8dc-4269-980f-f26feaf8f758",
  "model": "luma:ray@3.2",
  "positivePrompt": "Create a polished five-second documentary b-roll shot for an editorial film about traditional bookbinding. Follow the four supplied keyframes in exact chronological order. Begin with the artisan holding the wide brass-backed comb just above the near edge of the marbling tray. The hand lowers naturally and the steel tines enter the sizing, producing subtle concentric ripples without splashing. In one slow, continuous, physically plausible stroke, the artisan draws the comb from the near edge toward the far edge. As each tine passes through, the straight bands of indigo, saffron, oxblood red, muted teal, and cream pigment curl into a precise repeating peacock-chevron pattern behind the comb, while the untouched bands ahead remain straight. End as the comb reaches the far edge and lifts away; a few droplets fall back and the completed pattern gently settles. Preserve the same hand, charcoal linen sleeve, comb, tray, pigment colors, workshop, and lighting throughout. Use a close overhead three-quarter view with a very subtle slow push-in, stable framing, realistic liquid dynamics, natural hand speed, crisp tactile detail, warm window highlights, soft shadow falloff, restrained cinematic contrast, and premium photorealistic documentary stock-footage quality. No cuts, no time lapse, no extra hands, no paper entering the tray, no object morphing, no spilled paint, no text, and no logos.",
  "resolution": "720p",
  "duration": 5,
  "inputs": {
    "frameImages": [
      {
        "image": "https://assets.runware.ai/assets/inputs/e7571daa-49b0-47d4-a4d6-c3fa400500e0.jpg",
        "frame": 0
      },
      {
        "image": "https://assets.runware.ai/assets/inputs/00079cad-acb7-42ae-a9b3-3e806e9db643.jpg",
        "frame": 40
      },
      {
        "image": "https://assets.runware.ai/assets/inputs/90db2c44-0d60-4b4d-8b1d-d25fcf03080b.jpg",
        "frame": 80
      },
      {
        "image": "https://assets.runware.ai/assets/inputs/274c2b47-c26e-4cc0-a68a-abf86fc809b6.jpg",
        "frame": 120
      }
    ]
  }
}
```

---

### video-edit (video-edit)

[Watch video](https://assets.runware.ai/examples/luma-ray3-2/6052822e-7cc3-4aa2-a360-079836edd467.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "8a8ca742-c055-448f-a9d1-2f67d15b697a",
  "model": "luma:ray@3.2",
  "positivePrompt": "Transform the source footage into a premium documentary-style product film set inside a mobile elephant conservation clinic at blue hour. Preserve the source video's exact composition, camera move, timing, veterinarian performance, hand positions, device geometry, charcoal finish, turquoise probe cable, and every stage of the product interaction. Replace the plain studio with an open-sided khaki field tent in an East African wildlife reserve. Replace the silicone training phantom beneath the moving probe with the softly lit flank of a calm adult elephant standing beside the examination table, aligned precisely so the original probe contact and sweeping motion remain physically convincing. The veterinarian now wears practical sand-colored conservation field scrubs and thin medical gloves while retaining the same body shape and movement.\n\nUse cool indigo predawn ambience outside the tent, warm battery-powered work lamps within, subtle dust in the air, canvas seams, rugged equipment cases, folded blankets, and soft silhouettes of acacia trees in the distant background. Keep the ultrasound unit as the visual hero: crisp, clean, realistically integrated, consistent in scale, and illuminated with a controlled warm edge light. Add natural reflections and contact shadows appropriate to the new environment without altering the product's recognizable design. The elephant remains relaxed and nearly still apart from gentle breathing and a slight ear movement. Restrained natural color palette of indigo, khaki, charcoal, muted olive, and warm amber. Authentic conservation medicine rather than adventure spectacle. Cinematic but credible, shallow depth of field, refined commercial contrast, realistic skin and fabric textures, no visible branding, captions, interface overlays, or on-screen text.",
  "resolution": "720p",
  "settings": {
    "edit": {
      "autoControls": true
    }
  },
  "inputs": {
    "video": "https://assets.runware.ai/assets/inputs/47de5af9-c648-46e1-be09-5d46443a3b91.mp4"
  }
}
```

---

### video-reframe (video-reframe)

[Watch video](https://assets.runware.ai/examples/luma-ray3-2/f2d0d335-2bec-45a5-b67f-f3b22f538b3b.mp4)

```json
{
  "taskType": "videoInference",
  "taskUUID": "d708541e-9f06-417b-b93c-31bc2b977e29",
  "model": "luma:ray@3.2",
  "positivePrompt": "Create a museum-grade archival frame restoration by extending the source scene vertically onto a 9:16 portrait canvas. Keep every pixel of the original footage unchanged: preserve the technician, star projector, performance, camera position, timing, color, exposure, grain, dust and existing crop exactly as supplied. Reconstruct only the missing image area surrounding it. Continue the pale cyan planetarium dome upward with period-correct acoustic seams, subtle projector-cast reflections and softly falling illumination; continue the terrazzo floor, projector pedestal and orderly rows of dark 1960s theater seats downward. Match the source film's faded color-reversal palette, modest contrast, optical softness, fine grain and slight gate weave so the new margins feel like recovered portions of the same film reel. Maintain correct architecture, scale, perspective and coherent lighting across the joins. The completed portrait composition should center the technician and towering mechanical star projector with generous dome above and floor context below. Seamless reconstruction with no visible border, matte, split, duplicated objects, modern fixtures, added people, titles or graphics.",
  "width": 720,
  "height": 1280,
  "inputs": {
    "video": "https://assets.runware.ai/assets/inputs/3dcc5e4e-fd70-4436-8c3d-153bdd1d4a5b.mp4"
  }
}
```