Kling/kling-video-v4-0-reference-to-video/Released Oct 2026

Kling 4.0 - Reference to Video

Multi-reference video generation and editing

Commercial use

Video

About Kling 4.0 - Reference to Video

Kling 4.0 generates and edits video guided by up to 15 references: images, videos, and consistent character or object elements, with native audio in single takes of 3 to 30 seconds at 720p or 1080p. It keeps faces, outfits, and products consistent with the references while following the motion and camera of a reference video.

Add a video to edit and describe what to change and what to keep: swap an object, restyle a scene, or rewrite a line of dialogue, and the edit runs at the source video's length by default. Refer to inputs in the prompt as @image_1, @video_1, and @element_1, and say how each should be used, such as "keep the face and outfit from @image_1" or "follow the dance moves in @video_2 but not its background".

See what this model can do

A handful of real prompts and their outputs.

Character Reference

Bamboo Forest Discovery

"Keeps the character's hair, cloak and satchel from a single design reference."

Video Edit

Dawn to Bioluminescent Night

"Changes the light and the sea while every hatchling and wave keeps its original motion."

Product Reference

Picnic Pour

"A product ad that keeps the bottle and its label consistent with the reference."

Person and Place

The Clockmaker's Shop

"Combines a portrait reference and a location reference into one scene."

Style Reference

Watercolor Canal

"Follows the washes and paper texture of a watercolor reference while the scene moves."

Outfit Reference

Silver Puffer Run

"Puts the jacket from a product photo onto the person from a portrait reference."

Ready to build with Kling 4.0 - Reference to Video?

Try Kling 4.0 - Reference to Video in the Workbench to prompt it, compare outputs, and iterate on prompts without writing any code. When you're ready to ship, call the same model from our API and build your own apps on top of it.

curl -sSf -X POST https://hub.oxen.ai/api/ai/videos/generate \
    -H "Content-Type: application/json" \
    -H "Authorization: Bearer $OXEN_API_KEY" \
    -d '{
  "model": "kling-video-v4-0-reference-to-video",
  "prompt": "A cat sitting on a chair",
  "aspect_ratio": "adaptive",
  "resolution": "1080p",
  "duration": "auto"
}'

Request parameters

Every field Kling 4.0 - Reference to Video accepts in the request body. See the API reference for response formats.

POSThttps://hub.oxen.ai/api/ai/videos/generate

Field	Type	Required	Default	Description
`prompt`	`string`	Yes	—	Describe what to generate or change, and how to use each reference: @image_1, @video_1, @element_1, and so on.
`input_images`	`array<string>`	No	—	Reference images for characters, objects, scenes, or style. Up to 10 images in total, including first and last frames.
`input_video`	`string` uri	No	—	Video to edit. Describe what to change and what to keep.
`input_videos`	`array<string>`	No	—	Videos whose motion, camera, or style guide the output. Up to 5 videos including the video to edit.
`elements`	`array<object>`	No	—	Characters or objects to keep consistent, each from a front view plus other angles. Up to 7 elements.
`input_image`	`string` uri	No	—	Optional image the video starts on.
`tail_image_url`	`string` uri	No	—	Optional image the video ends on.
`aspect_ratio`	`string`	No	`"adaptive"`	`adaptive16:99:161:121:9`
`resolution`	`string`	No	`"1080p"`	`720p1080p`
`duration`	`integer \| string`	No	`"auto"`	Range: 3 – 30

Model JSON Video generation reference Async queue Models endpoint