kling
Generate video and images with Kling AI models via inference.sh CLI. Models: Kling Video V4, V4 Flash, O1 (Omni), V3, Kling Avatar, Lip Sync, Video-to-Audio, Image V3, Image 3O, Virtual Try-On. Capabilities: text-to-video, image-to-video with start and end frames, 3-30s duration, native audio, automatic shot-splitting, reference-image and reference-video editing, talking avatars, lip sync, sound effects and music for video, text-to-image up to 4K, clothing try-on. Use for: cinematic clips, product videos, social media, dubbing, digital presenters, fashion try-on. Triggers: kling, kling ai, kli
Try your first task
A source-grounded starter task has not been verified for this skill yet.
SKILL.md
name: klingdescription: "Generate video and images with Kling AI models via inference.sh CLI. Models: Kling Video V4, V4 Flash, O1 (Omni), V3, Kling Avatar, Lip Sync, Video-to-Audio, Image V3, Image 3O, Virtual Try-On. Capabilities: text-to-video, image-to-video with start and end frames, 3-30s duration, native audio, automatic shot-splitting, reference-image and reference-video editing, talking avatars, lip sync, sound effects and music for video, text-to-image up to 4K, clothing try-on. Use for: cinematic clips, product videos, social media, dubbing, digital presenters, fashion try-on. Triggers: kling, kling ai, kling 4, kling v4, kling 4.0, kling flash, kling o1, kling omni, kling 3, kling v3, kling avatar, kling lip sync, kling lipsync, kolors, kling image, kling try-on, virtual try-on, kuaishou video"allowed-tools: Bash(belt *)Install the belt CLI skill:
npx skills add belt-sh/cli
Kling Video and Image Generation
Generate video, avatars, lip sync, audio and images with Kling AI models via inference.sh CLI.
Quick Start
Requires inference.sh CLI (
belt). Install instructions
belt login
belt app run klingai/video-v4 --input '{"prompt": "a red fox trotting through fresh snow at dawn"}'
Models
| Model | App ID | Best For |
|---|---|---|
| Kling Video V4 | klingai/video-v4 | Text/image-to-video, start and end frames, 3-30s, 720p/1080p, native audio |
| Kling Video V4 Flash | klingai/video-v4-flash | Faster, cheaper V4. 720p, 3-20s, first-frame image-to-video |
| Kling Video O1 | klingai/video-o1 | Omni model: reference images, reference video for style transfer or editing |
| Kling Video V3 | klingai/video-v3 | Up to 4K, multi-shot, 3-15s, synchronized audio |
| Kling Avatar | klingai/avatar | Talking head video from one face photo and audio |
| Kling Lip Sync | klingai/lip-sync | Drive mouth movement in an existing video with audio |
| Kling Video to Audio | klingai/video-to-audio | Add sound effects and background music to a 3-20s video |
| Kling Image V3 | klingai/image-v3 | Text-to-image at 1K/2K |
| Kling Image 3O | klingai/image-3o | Text-to-image and image-to-image up to 4K, series images |
| Kling Virtual Try-On | klingai/virtual-tryon | Put a clothing item on a person photo |
V4 and V4 Flash bill per second and split shots automatically. A duration field takes precedence over a duration named in the prompt.
Older versions (Video V2.5, V2.6, Image V1 to V2.1, Image O1) are still available:
belt app search "kling"
Examples
Text-to-Video (V4)
belt app run klingai/video-v4 --input '{
"prompt": "a chef plating a dessert in a busy restaurant kitchen, close-up, steam rising",
"duration": 10,
"resolution": "1080p",
"aspect_ratio": "16:9"
}'
Image-to-Video with Start and End Frames (V4)
belt app run klingai/video-v4 --input '{
"prompt": "the flower slowly blooms as morning light hits it",
"image": "https://closed-bud.jpg",
"end_image": "https://open-flower.jpg",
"duration": 5
}'
With an image, the video follows the image aspect ratio; aspect_ratio applies to text-to-video only.
Fast Vertical Clip (V4 Flash)
belt app run klingai/video-v4-flash --input '{
"prompt": "a skateboarder kickflips down a stair set, handheld camera",
"aspect_ratio": "9:16",
"duration": 8
}'
Reference Images (O1)
Reference inputs in the prompt as @image_1, @image_2. Up to 7 reference images without a video, 4 with one.
belt app run klingai/video-o1 --input '{
"prompt": "@image_1 walks through the market from @image_2, evening light",
"reference_images": ["https://character.jpg", "https://market.jpg"],
"aspect_ratio": "16:9",
"duration": 8
}'
Edit a Video (O1)
reference_video_type is feature (use the video as a style/motion reference) or base (edit the video directly).
belt app run klingai/video-o1 --input '{
"prompt": "change the season in @video_1 to winter with falling snow",
"reference_video": "https://your-video.mp4",
"reference_video_type": "base"
}'
4K Multi-Shot (V3)
belt app run klingai/video-v3 --input '{
"prompt": "a lighthouse on a cliff during a storm, waves crashing, lightning",
"resolution": "4k",
"duration": 10,
"sound": true,
"multi_shot": true
}'
Talking Avatar
belt app run klingai/avatar --input '{
"image": "https://face-photo.jpg",
"audio": "https://speech.mp3",
"prompt": "while talking, nods and smiles warmly",
"mode": "pro",
"aspect_ratio": "9:16"
}'
Lip Sync an Existing Video
belt app run klingai/lip-sync --input '{
"video": "https://speaker.mp4",
"audio": "https://new-dialogue.mp3",
"original_audio_volume": 0
}'
Add Sound to a Video
belt app run klingai/video-to-audio --input '{
"video": "https://beach-clip.mp4",
"sound_effect_prompt": "ocean waves, seagulls, gentle wind",
"bgm_prompt": "calm acoustic guitar"
}'
Images
# Image V3, 2K
belt app run klingai/image-v3 --input '{
"prompt": "minimalist product shot of a ceramic mug on linen, soft daylight",
"resolution": "2k",
"aspect_ratio": "4:3",
"n": 2
}'
# Image 3O, 4K with a reference image
belt app run klingai/image-3o --input '{
"prompt": "the same sneaker on a wet city street at night, neon reflections",
"image": "https://sneaker.jpg",
"resolution": "4k",
"aspect_ratio": "3:2"
}'
Virtual Try-On
belt app run klingai/virtual-tryon --input '{
"human_image": "https://person-full-body.jpg",
"cloth_image": "https://jacket-product.jpg"
}'
V1.5 (default, kolors-virtual-try-on-v1-5) accepts upper and lower garments merged into one clothing image.
Workflow: Image to Video with Sound
# 1. Generate a keyframe
belt app run klingai/image-v3 --input '{"prompt": "astronaut standing on a red desert planet, wide shot", "aspect_ratio": "16:9"}'
# 2. Animate it with V4
belt app run klingai/video-v4 --input '{"prompt": "the astronaut starts walking toward the horizon, dust blowing", "image": "<image-url-from-step-1>"}'
Parameters (Video V4)
| Parameter | Type | Default | Description |
|---|---|---|---|
prompt | string | required | Text description of the video |
image | file | - | Start frame (jpg/png, max 30MB, min 300px per side, aspect ratio 2:5 to 5:2) |
end_image | file | - | End frame, requires image |
resolution | enum | 1080p | 720p, 1080p |
aspect_ratio | enum | 16:9 | 21:9, 16:9, 9:16, 1:1 (text-to-video only) |
duration | integer | 5 | 3-30 seconds, billed per second |
V4 Flash takes prompt, image, aspect_ratio (21:9, 16:9, 9:16, 1:1) and duration (3-20). It has no end_image or resolution and outputs 720p.
Parameters (Video O1)
| Parameter | Type | Default | Description |
|---|---|---|---|
prompt | string | required | Use @image_1, @video_1 to reference inputs. Max 2500 chars |
image / end_image | file | - | First and end frame |
reference_images | file[] | [] | Up to 7 (4 with a reference video) |
reference_video | file | - | Video for style/motion reference or editing |
reference_video_type | string | feature | feature or base |
resolution | enum | 1080p | 720p, 1080p |
aspect_ratio | enum | - | 16:9, 9:16, 1:1. Required for text-to-video |
duration | integer | 5 | Text-to-video: 5 or 10. With reference images: 3-10 |
watermark | boolean | false | Add watermark |
Parameters (Video V3)
| Parameter | Type | Default | Description |
|---|---|---|---|
prompt | string | required | Max 3072 chars (2500 recommended) |
image / end_image | file | - | Start and end frame |
sound | boolean | true | Generate synchronized audio |
multi_shot | boolean | true | Multi-shot generation via prompt format |
resolution | enum | 1080p | 720p, 1080p, 4k |
aspect_ratio | enum | 16:9 | 16:9, 9:16, 1:1 |
duration | integer | 5 | 3-15 seconds |
Pricing
| App | Price |
|---|---|
klingai/video-v4 | $0.157/s at 720p, $0.235/s at 1080p |
klingai/video-v4-flash | $0.094/s at 720p |
klingai/avatar | $0.056/s (std), $0.112/s (pro) |
klingai/lip-sync | $0.014/s, 5s minimum |
klingai/video-to-audio | $0.0088 per video |
klingai/image-v3 | $0.0035 per image |
klingai/image-3o | $0.028 per image (1K/2K), $0.056 (4K) |
For O1, V3 and Try-On, check belt app pricing <app-id>. Estimate a run before paying: belt app estimate <app-id> --input input.json.
Sample Workflow
# 1. Generate sample input to see all options
belt app sample klingai/video-v4 --save input.json
# 2. Edit the prompt
# 3. Run
belt app run klingai/video-v4 --input input.json
Related Skills
# Full platform skill (all apps)
npx skills add inference-sh/skills@infsh-cli
# All video generation models
npx skills add inference-sh/skills@ai-video-generation
# Image-to-video workflows
npx skills add inference-sh/skills@image-to-video
# AI avatars & lipsync
npx skills add inference-sh/skills@ai-avatar-video
# Video prompting
npx skills add inference-sh/skills@video-prompting-guide
# Other video models
npx skills add inference-sh/skills@seedance
npx skills add inference-sh/skills@google-veo
npx skills add inference-sh/skills@runway
npx skills add inference-sh/skills@pixverse
Browse all video apps: belt app list --category video
Documentation
- Running Apps - How to run apps via CLI
- Streaming Results - Real-time progress updates
- Content Pipeline Example - Building media workflows