# Kling v3.0 Motion Control

Kling v3.0 Motion Control is the v3.0 motion transfer model in the Kling series. You supply a reference image that defines your character and a reference video that defines the motion. Kling v3.0 Motion Control combines the two into a new clip where facial identity stays stable through complex, long-duration movement.

- **Model ID:** `klingai/kling-v3.0-motion-control`
- **Type:** video
- **Providers:** klingai
- **Context window:** Not applicable
- **Maximum output tokens:** 0
- **Pricing:** See the provider details below
- **Canonical page:** https://vercel.com/ai-gateway/models/kling-v3.0-motion-control

## Supported parameters

- **Parameters:** `aspectRatio`, `resolution`, `duration`, `fps`
- **Operations:** `image-to-video`
- **Input modalities:** `text`, `image`, `video`
- **Output modalities:** `video`
- **Aspect ratios:** `16:9`, `9:16`, `1:1`
- **Resolutions:** `720p`, `1080p`
- **Durations:** `3s`, `4s`, `5s`, `6s`, `7s`, `8s`, `9s`, `10s`, `11s`, `12s`, `13s`, `14s`, `15s`, `16s`, `17s`, `18s`, `19s`, `20s`, `21s`, `22s`, `23s`, `24s`, `25s`, `26s`, `27s`, `28s`, `29s`, `30s`
- **Frame rates:** `30 fps`
- **Maximum outputs per request:** Not reported
- **Maximum input images:** 1
- **Input image formats:** `jpg`, `jpeg`, `png`
- **Maximum input image size:** 10 MB

## Example

```ts
import { experimental_generateVideo as generateVideo } from 'ai';

const result = await generateVideo({
  model: 'klingai/kling-v3.0-motion-control',
  prompt: 'A serene mountain lake at sunrise.'
});
```

## About

Kling v3.0 Motion Control generates video where character actions match a reference video while visual appearance comes from a reference image. You provide a still image of the character, a clip containing the motion to capture, and an optional text prompt that guides elements, backgrounds, and motion effects. Kling v3.0 Motion Control maps full-body posture, joint movement, hand gestures, and facial expressions from the clip onto your character.

Character fidelity is the focus of the v3.0 generation. Facial features stay stable across multi-angle and long-duration motion, and subtle emotional transitions (smiling, surprise, sadness) carry over accurately. When hands, hats, props, or fans partially cover the face, Kling v3.0 Motion Control restores facial detail from the reference imagery across frames. Clarity also holds while the camera zooms, pans, or tracks the subject, so motion transfer combines cleanly with camera direction in the prompt.

Two character orientation modes control framing and duration. In image mode the character keeps facing the direction shown in the still, with output up to 10 seconds. In video mode the character's orientation follows the reference clip, with output up to 30 seconds. Output renders at 720p in std mode or 1080p in pro mode, and this page lists the live per-second rates for each.

Compared with Kling v2.6 Motion Control, the v3.0 release improves element consistency, preserves character identity more reliably, and produces smoother motion transfer. Choose Kling v3.0 Motion Control when a specific character must perform a specific referenced motion and identity drift is not acceptable.

## What to consider

Reference quality drives transfer accuracy. Keep the character's full body and head clearly visible in the reference image, and match proportions between the image and the video. Avoid pairing a full-body reference clip with a half-body image. Steady, clear movements transfer best. Very fast or chaotic motion can degrade results.

Video generation is in beta for Pro and Enterprise plans and paid AI Gateway users. Confirm your plan before you ship.

## When to use

### Best For
- Character motion transfer where a reference performance applies to a character defined by a single still image
- Dance and performance clips replicating body movement, gestures, and timing on a new character or scene
- Virtual avatars and brand mascots that must keep facial identity stable while performing referenced motion
- Motion transfer combined with camera direction like tracking shots, pans, and zooms

### Consider Alternatives When
- You don't have a reference motion video and prefer purely text-driven generation. Use v3.0 t2v
- You want to animate a still image without motion transfer. Use v3.0 i2v
- Speed and cost matter more than motion fidelity. Kling's turbo tiers are faster and cheaper

## Best for

- **Character Motion Transfer:** Applying a reference performance to a character defined by a single still image
- **Dance and Performance Clips:** Replicating body movement, gestures, and timing on a new character or scene
- **Virtual Avatars and Mascots:** Keeping facial identity stable while a brand character performs referenced motion
- **Camera-Aware Staging:** Combining motion transfer with tracking shots, pans, and zooms without losing clarity

## Consider alternatives

- **No Reference Video:** You prefer purely text-driven generation without a motion clip, so use v3.0 t2v
- **Simple Image Animation:** You want to animate a still image without motion transfer, so use v3.0 i2v
- **Speed and Cost Priority:** Generation speed and cost matter more than motion fidelity, so Kling's turbo tiers are faster and cheaper

## Frequently asked questions

### What inputs does Kling v3.0 Motion Control require?

Two inputs are required: a reference image that defines the character's appearance and a reference video that defines the motion. An optional text prompt guides elements, backgrounds, and motion effects. Keep the character's full body and head clearly visible in the image.

### How do the character orientation modes affect output duration?

Image mode keeps the character facing the direction shown in the reference image and caps output at 10 seconds. Video mode follows the orientation in the reference clip and extends output up to 30 seconds.

### What resolutions does Kling v3.0 Motion Control support?

Output renders at 720p in std mode and 1080p in pro mode. Std mode is the cost-effective tier; pro mode targets higher quality. This page lists the live per-second rate for each mode.

### How does Kling v3.0 Motion Control keep facial identity stable during occlusions?

When hands, hats, props, or fans partially cover the face, Kling v3.0 Motion Control restores facial detail from the reference imagery across frames. Facial features also stay stable across multi-angle and long-duration motion, and clarity holds while the camera zooms, pans, or tracks.

### How does v3.0 Motion Control differ from v2.6 Motion Control?

V3.0 improves element consistency, preserves character identity more reliably, and produces smoother motion transfer than v2.6. V3.0 also centers the workflow on a character reference image, so a defined subject performs the referenced motion with stable facial identity.

### How do I use Kling v3.0 Motion Control through AI Gateway?

Call Kling v3.0 Motion Control with `generateVideo` from the AI SDK, passing your reference image, reference video, and optional prompt. Video generation is in beta for Pro and Enterprise plans and paid AI Gateway users, so confirm your plan before you ship.

## Links

- [Model page](https://vercel.com/ai-gateway/models/kling-v3.0-motion-control)
- [AI Gateway documentation](https://vercel.com/docs/ai-gateway)
- [Provider model documentation](https://app.klingai.com/global/dev/document-api/apiReference/model/motionControl)
- [Provider pricing](https://klingai.com/global/dev/pricing?scrollTo=video)