Kling 4.0 Reference to Video - ComfyUI Built-in Node Documentation

Generates a video from reference images and reference videos with the Kling 4.0 model, including audio. The references drive the subject, style and motion, and an optional first and last frame pin the start and the end of the clip. Use it to keep a character or a scene consistent across a shot while describing the action in the prompt.

Inputs

Common Inputs

Parameter	Description	Data Type	Required	Range
`model`	Model selector. Only `"kling-4.0"` is available; selecting it reveals the inputs below.	DYNAMIC_COMBO	Yes	`"kling-4.0"`

kling-4.0 Inputs

These inputs appear when model is set to "kling-4.0".

Parameter	Description	Data Type	Required	Range
`reference_images`	Growable slots for reference images, up to 10. Refer to them in the prompt as `@Image1`, `@Image2` and so on, in connection order.	AUTOGROW	No	0 to 10 images
`reference_videos`	Growable slots for reference videos, up to 5. Refer to them in the prompt as `@Video1`, `@Video2` and so on. Each video and all videos together must be 3 to 30 seconds long.	AUTOGROW	No	0 to 5 videos
`first_frame`	Optional frame the video starts on. Counts toward the 10-image limit together with the reference images.	IMAGE	No	-
`last_frame`	Optional frame the video ends on. Counts toward the 10-image limit together with the reference images.	IMAGE	No	-
`prompt`	Description of the video to generate. Refer to the inputs as `@Image1`, `@Image2` and so on, and `@Video1`, `@Video2` and so on.	STRING	Yes	1 to 26000 characters
`resolution`	Resolution of the output video. Default: `"720p"`.	COMBO	Yes	`"720p"` `"1080p"`
`aspect_ratio`	Aspect ratio of the output video. `"adaptive"` follows the prompt, then the first frame or the reference videos, and otherwise falls back to 16:9. Default: `"adaptive"`.	COMBO	Yes	`"adaptive"` `"16:9"` `"9:16"` `"1:1"` `"21:9"`
`duration`	Length of the generated video in seconds. Default: 5.	INT	Yes	3 to 30
`seed`	Seed controls whether the node should re-run; results are non-deterministic regardless of seed. Default: 42.	INT	Yes	0 to 2147483647

Constraints and behavior:

Images are limited to 10 in total, counting the reference images, the first frame and the last frame.
Every image must be at least 300x300 pixels with an aspect ratio between 2:5 and 5:2.
Every reference video must be 3 to 30 seconds long, at least 700x700 and at most 4553x4553 pixels, no more than 8,294,400 pixels in total, with an aspect ratio between 2:5 and 5:2 and a frame rate of at least 24 fps. All reference videos together must not exceed 30 seconds.
A prompt that mentions @ImageN or @VideoN raises an error when fewer than N inputs of that kind are connected.
When a first or a last frame is connected, the node adds an instruction that the video starts or ends exactly on that frame.

Outputs

Output Name	Description	Data Type
`video`	The generated video, downloaded from the provider once the task completes.	VIDEO

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): bda2298ed6ed06db04a8f342ec6ad0d5a0494e09bfadee72fce4420ebe4eaf6a

​Inputs

​Common Inputs

​kling-4.0 Inputs

​Outputs

Inputs

Common Inputs

kling-4.0 Inputs

Outputs