Transfer motion from any reference video to a static image with preserved identity and smooth animation








Kling 3.0 Motion Control extracts motion from a reference video and applies it to a static image. Upload a photo and a video — the model maps the movement onto your character while preserving their appearance. Output ranges from 3 to 30 seconds depending on settings.
Extract walking, dancing, gestures, and choreography from any reference video and apply it to your image.
The character in your image keeps their face, clothing, and proportions throughout the animation.
Output up to 10 seconds with image orientation or up to 30 seconds with video orientation.
Better element consistency, smoother motion transfer, and higher output quality compared to the 2.6 version.
Provide a clear, well-lit photo of your character. JPG or PNG, max 10MB, 340-3850px. Full body shots work best.
Add the video containing the motion you want to transfer. MP4 or MOV, max 100MB, 3-30 seconds.
Choose 'image' to prioritize your character's appearance (max 10s) or 'video' to follow the reference motion more closely (max 30s).
Optionally add a prompt to guide elements, background, and actions. Select Standard or Pro mode and generate.
Captures walking, running, dancing, hand gestures, facial expressions, and complex choreography sequences.
Set to 'image' to keep your character's identity dominant, or 'video' to follow the reference motion more faithfully.
Standard mode for faster output. Pro mode for higher quality with better detail preservation.
Keep the audio from your reference video in the final output.
Add text prompts to control background, lighting, and additional scene elements beyond the motion transfer.
Generate up to 30 seconds of motion-controlled video in video orientation mode.
Credits are based on video length and quality mode. The generator shows the exact estimate before you create.
5-second video: 75 credits. 10-second video: 150 credits.
5-second video: 150 credits. 10-second video: 300 credits.
Best for transferring specific movements from video to a character image.
Transfer dance routines from reference videos to any character — real or illustrated.
Make static photos come alive with trending dance moves or gestures for TikTok, Reels, and Shorts.
Animate brand characters or mascots with natural human movement without rigging or motion capture equipment.
Preview choreography or movement sequences on different characters before committing to live production.
Transfer motion from any video to your character image.
Kling 4.0 is coming soon for 4K+ cinematic AI video from text and images. Native audio, multi-shot sequencing, persistent character identity, and enhanced photorealism are expected in a single generation workflow.
Generate native 4K AI videos with Kling 3.0. Multi-shot sequencing, integrated audio generation, text-to-video and image-to-video — all in a single generation workflow.
Generate and edit AI videos from text, images, and video references with Kling 3.0 Omni. Reference-based character consistency, video-to-video editing, and native audio in one unified model.
Generate fast, affordable AI videos with Kling O3. Text-to-video, image-to-video, multi-shot sequencing, native audio, and 4K output — at a lower credit cost than Kling 3.0.
Turn any portrait photo into a talking video with Kling Avatar V2. Upload a face image and an audio file — the model generates precise lip sync, natural head motion, and facial expressions at 1080p 48fps.
Generate cinematic AI videos with Kling 2.6. Native audio, accurate lip sync, 1080p output, 5s or 10s duration. The most affordable Kling model for single-shot video with sound.
Control how elements move in your video — paint paths, transfer motion from reference clips, animate up to 6 elements
Generate and edit high-quality AI images with Kling O3. Text-to-image generation and image editing with reference inputs — 1K to 4K resolution, multiple aspect ratios, 5 credits per image.
Generate ultra-fast photorealistic AI images with Nano Banana 2. Text-to-image and image-to-image generation in 1K, 2K, or 4K resolution across a wide range of aspect ratios.