Kling 2.6 Motion Control on Genosai — character motion from a reference video

Kling 2.6 Motion Control is the motion control mode in the Kling line by Kuaishou. You upload a reference video that sets the movement trajectory and poses, and the model transfers that choreography onto your character while keeping its appearance. On Genosai clips are available at 720p and 1080p from 6 credits per second, online and with no API keys.

Updated: July 7, 2026

• Motion transfer from a sample — A reference video sets the trajectory and poses, and the model repeats that choreography on your character.
• Identity preserved — The character keeps its face and appearance while fully inheriting the motion from the sample.
• Predictable choreography — Control over the camera trajectory and character path gives a repeatable, choreographed result.
• 720p and 1080p — Pick the resolution for the task — the price is counted per second of the clip, not per fixed clip.
• Precise motion trajectory — Motion is set by a sample, not a random reading of text, so the result is repeatable.

Contents

• What is Kling 2.6 Motion Control
• Capabilities
• Examples
• How to use on Genosai
• Prompts
• Generation cost
• How it compares
• Limitations and tips
• FAQ

What is Kling 2.6 Motion Control

Kling 2.6 Motion Control is a video generation mode by Kuaishou (Kling AI) built around motion transfer. Unlike ordinary generation, where the model invents the motion from text, here the movement trajectory is taken from a ready reference video and transferred onto your character. On Genosai the mode is available online, so you do not need API keys or your own infrastructure — generation starts right in the browser.

How it works in essence. You give the model two inputs: the character likeness and a reference video with the choreography you need. The model binds the movement trajectory and poses from the sample to your subject, keeps its appearance, and returns a clip where the character repeats the given motion. That is a fundamentally different control than a text prompt: instead of describing the motion in words, you show it by example.

Why this matters. A text prompt sets the plot well but describes precise choreography poorly — a specific dance, gesture or camera path. Motion Control solves that: the motion comes out exactly as in the sample, and it can be repeated frame to frame. As a result the outcome is predictable rather than changing randomly from one generation to the next.

Who is Kling 2.6 Motion Control for first of all. It is authors of dance and music clips who need a specific choreography. It is creators of avatars and talking characters who need repeatable motion. It is designers assembling scenes with a pre-planned camera trajectory. For scenes with sound without motion transfer, see regular Kling 2.6, and for more advanced transfer, Kling 3.0 Motion Control.

Capabilities

Kling 2.6 Motion Control covers tasks where motion must be set precisely and repeatably. Below is what the mode does confidently in practice.

Motion transfer from a sample

The core of the mode is motion transfer. The model takes the movement trajectory and the sequence of poses from the reference video and transfers them onto your character. The character inherits the sample's choreography but stays itself: face and appearance are kept. This is handy when you have a motion reference — a dance or gesture recording, for example — to repeat on a different subject.

Control over camera and character trajectory

Beyond body motion, the mode gives spatial and temporal control: the camera trajectory and the character path in frame. As a result the scene comes out choreographed and predictable — you know in advance how the motion will go rather than relying on a random reading of the text.

Character identity preserved

During transfer the model holds the subject's appearance: facial features and overall look do not drift from frame to frame. The character fully inherits the sample's dynamics while staying recognizable. This is key for avatars and serial clips where the subject must look the same.

Resolution choice

Clips are available at 720p and 1080p. You choose the resolution for the task, and the price is counted per second of the clip — so spending scales with duration rather than being tied to a fixed clip. For motion drafts it is handy to use 720p and regenerate the final variant at 1080p.

Examples

There are no demo clips on the page yet — video generation is expensive, and we will add them later. The model cover image, however, shows a result typical of Kling 2.6 Motion Control. Below are the typical scenarios teams use the mode for:

• A dance clip: a character repeats the choreography from an uploaded reference video while keeping its appearance.
• An avatar with a gesture: a talking or gesturing subject repeats the motion of a reference recording.
• Repeating a walk or action: transferring a specific body motion onto a new character for a series of clips.
• Camera choreography: a pre-set camera trajectory around the character for a predictable scene.
• Styling to a sample: the same dance or gesture on different subjects in one unified motion.

Finished clips can be downloaded and used in videos and stories. For scenes where motion is set by text rather than a sample, see regular Kling 2.5 Turbo.

How to use on Genosai

Launching Kling 2.6 Motion Control on Genosai needs no technical setup: everything happens in the browser and the mode is already connected. Below is the basic path from sign-in to a finished clip.

1. Sign in to your Genosai account and open the video studio.
2. Pick the Kling 2.6 Motion Control mode.
3. Upload the likeness of the character whose motion you want to set.
4. Upload the reference video with the choreography to repeat.
5. Set the resolution (720p or 1080p) and the clip duration.
6. Send the request, refine the scene and surroundings with a prompt if needed, then regenerate a strong variant.

Prompts

In Motion Control the reference video does the main work, but a text prompt helps set the scene, style and surroundings. These templates show how to complement motion transfer with description. Replace the text in brackets with your own.

Character [description] repeats the motion from the sample, studio background, even light, keep the subject's appearance

A dancer in [style] on stage, choreography from the reference video, spotlight, dynamic camera

Avatar [description] gestures per the sample, clean background, soft light, precise face preservation

Character [description] follows the reference motion, urban background, golden hour, keep identity across frames

Character walks through [location], gait from the sample, smooth side camera trajectory, cinematic light

Dance clip: transfer the motion from the uploaded video onto the subject [description], neon background, 1080p

Character performs [action] exactly per the sample, neutral background, focus on the motion, stabilized camera

Generation cost

On Genosai Kling 2.6 Motion Control is priced per second of the clip and by resolution — you pay for the actual generation, so a short clip costs less than a long one. The reference points: 720p costs 6 credits per second, 1080p costs 9 credits per second. For example, a 5 second clip at 720p costs about 30 credits, and at 1080p about 45. You choose resolution and duration before running, so spending is always predictable.

A practical approach is to debug the motion transfer on short takes at 720p, then regenerate the final variant at 1080p with the same sample. Starter credits after sign-up let you try the mode for free, and top-ups are available. See current rates and balance in the Pricing section.

How it compares

Kling 2.6 Motion Control is the pick when motion must be set from a reference video. If choreography does not matter and you need text-based generation with sound, regular Kling 2.6 fits. More advanced transfer with face locking and physics comes from Kling 3.0 Motion Control. For fast silent clips from text, use Kling 2.5 Turbo, and for a fast photo-animation alternative look at Seedance 2.0 Fast.

All modes are available in one Genosai interface, so you can compare them on your task. A sensible tactic is to use Motion Control where a specific choreography matters, and ordinary generation for free scenes from text.

Limitations and tips

Kling 2.6 Motion Control is strong at motion transfer, but it has its limits. The result depends heavily on the quality of the reference video: if the sample has blurred or ambiguous motion, the transfer comes out worse. The mode needs two inputs — a character and a sample — so it is less handy for quick improvisation than single-text generation. We do not state exact native specs beyond the resolutions available on Genosai, so as not to present unconfirmed numbers as fact.

Another common trap is a poorly chosen sample. For a clean transfer, use a video with clear, readable choreography and a contrasting background. Upload the character frontally and without heavy occlusion: the model then holds the face and appearance more accurately across frames. Use the text prompt to set the background, light and scene style, not to describe the motion itself — the sample sets that.

For the best result, work iteratively: debug the transfer on short 720p takes, pick a strong variant and regenerate it at 1080p. And if you need more precise transfer with facial feature locking and believable physics, compare the result with Kling 3.0 Motion Control on the same sample.

FAQ

What is Kling 2.6 Motion Control?

Kling 2.6 Motion Control is a video generation mode by Kuaishou that transfers motion from a reference video onto your character. You upload a reference video with the choreography you need, and the model repeats that movement trajectory and poses while keeping the subject's appearance. On Genosai the mode is available online with no API keys.

How does Motion Control differ from regular Kling 2.6?

Regular Kling 2.6 generates motion from a text description, while Motion Control takes the trajectory from a ready reference video and transfers it onto the character. That gives a precise, repeatable result where a specific choreography matters rather than a random reading of the prompt.

What do I need to upload for a generation?

Motion Control needs two inputs: an image or likeness of the character and a reference video with the motion to repeat. The model binds the movement trajectory from the sample to your character and keeps its appearance across every frame. The quality of the sample directly affects the result.

How much does one generation cost?

On Genosai Kling 2.6 Motion Control is priced per second of the clip by resolution: 720p costs 6 credits per second, 1080p costs 9 credits. For example, a 5 second clip at 720p costs 30 credits. You pay for actual generations, with no subscription.

What tasks is Motion Control good for?

The mode is handy when you need to repeat a specific motion: a dance, a gesture, a walk or camera choreography on your own character. It is used for dance clips, avatars and scenes where predictable, repeatable dynamics matter instead of random motion.

Do I need API keys to use the model?

No. Genosai gives access to Kling 2.6 Motion Control right in the browser — no API keys or infrastructure setup. Just sign in to Genosai and pick the mode in the video studio.

Try Kling 2.6 Motion Control on Genosai