🇺🇸English
WaveSpeed Kling Motion Control
Transfer motion from reference videos to animate still images using the WaveSpeed AI Kling v2.6 Motion Control API.
Prerequisites
uv (Python package manager)WAVESPEED_API_KEY environment variable set with a valid WaveSpeed AI API key (get one at https://wavespeed.ai)
How It Works
Motion control transfers movement from a reference video to a still image:
- Provide a still image of the subject you want to animate
- Provide a reference video with the desired motion
- Choose character orientation to match the subject's direction to either the image or video
- The API animates the image using the motion from the video
Configuration Parameters
| Parameter | CLI Flag | Values |
|---|
| Character orientation | --orientation | image (default), video |
| Prompt | positional or --prompt | Positive prompt for generation |
| Negative prompt | --negative-prompt | Negative prompt to avoid unwanted elements |
| Keep audio | --keep-audio | Preserve original video audio (default: true) |
Usage
Run from the skill folder (skills/wavespeed-keling-motion/):
Basic motion transfer
uv run --project runtime runtime/generate_motion.py --image person.png --video dance.mp4
With prompt guidance
uv run --project runtime runtime/generate_motion.py "A person dancing gracefully" \
--image person.png --video dance.mp4
Character orientation from video
Use when the reference video's character direction should be preserved:
uv run --project runtime runtime/generate_motion.py --image face.png --video nod.mp4 \
--orientation video
With negative prompt
uv run --project runtime runtime/generate_motion.py "A robot walking" \
--image robot.png --video walk.mp4 \
--negative-prompt "blurry, distorted, low quality"
Discard original audio
uv run --project runtime runtime/generate_motion.py --image singer.png --video lip_sync.mp4 \
--no-keep-audio
Custom output
uv run --project runtime runtime/generate_motion.py --image cat.png --video stretch.mp4 \
--filename cat_stretch.mp4 -o path/to/output/
Output
- Default output directory:
output/ inside the skill folder.
- Output filenames:
motion_YYYYMMDD_HHMMSS.mp4 (timestamp-based) or custom via --filename.
- Videos are MP4 format.
Limitations
- Image inputs: max 10MB, formats: PNG, JPEG. Min 300px width/height. Aspect ratio 1:2.5 to 2.5:1.
- Video inputs: max 10MB, formats: MP4, MOV. Min 300px width/height. Aspect ratio 1:2.5 to 2.5:1.
- API requires top-up to activate. Keys without top-up will not work.
- Safety filters may block certain content.
Weekly Installs
1
Repository
anyt-io/notebook
GitHub Stars
9
First Seen
1 day ago
Security Audits
Gen Agent Trust HubPassSocketPassSnykPass
Installed on
mcpjam1
github-copilot1
junie1
windsurf1
zencoder1
crush1
