AI video model
Create motion-controlled video using image and video references with Kling 3.0 Motion Control.
Compare Kling 3.0 Motion Control with other video models — same workspace, one click to switch.
A glimpse of what people are creating with AI video models on CreateVision AI.
Kuaishou's motion-transfer model maps movement from a reference video onto your character image — preserving face, expression, and identity with physics communities call mathematically dependable.
Upload a character image and a motion reference video; Kling transfers the movement — dance, gesture, action — onto your character while preserving facial identity and expression. The dance-transfer and character-swap workflows that fill social feeds run on exactly this.
Full-body tracking with accurate hand and finger articulation — reviewers specifically note individual knuckle definition where earlier versions collapsed hands into mush. Choreography and sign-level gesture work become viable.
Fighting, hugging, object interaction — high-difficulty contact physics render with a consistency users describe as mathematically predictable. Complex action that breaks other models is this model's home turf.
Generate at 720p or 1080p with per-second pricing shown upfront. Identity-stable, physics-true motion at broadcast-friendly resolution makes it a practical tool for creator content at scale.
One account, the full image and video generation workflow.
Start free with daily credits. Upgrade for premium models, faster generation, and 4K output.
Perfect for getting started
Billed annually · $240/year
Billed annually · $300/year
It transfers movement from a reference video onto your character image — keeping the character's face, expression, and identity while they perform the reference motion.
At minimum a character image and a motion reference video. You can choose whether the character's orientation follows the image or the video.
Per second of output at 720p or 1080p rates, displayed before generation. Failed generations refund automatically.
Dance and performance transfers, character swaps, AI-influencer content, and action sequences with complex contact physics.
Yes — there is an option to keep the reference video's sound in the output.
It is tuned for realistic motion; highly stylized 2D animation transfers are less consistent. For cinematic scene generation without motion reference, use Seedance or Veo.
