Kling AI

Generate images and videos using Kling AI

Overview

The Kling AI tool interfaces with the Kling AI API for AI-powered media generation. It supports image generation from text prompts (with an optional reference image for image-to-image), text-to-video, image-to-video with an optional end frame, multi-image-to-video, extending an existing Kling video, single-image and dual-character video effects, lip sync onto an existing video (text-to-speech or an audio file), and virtual try-on (overlay clothing onto model images). All generation operations are asynchronous — tasks are submitted and can optionally be polled until completion, and the result media can be downloaded as binary data. Image generation and video generation use separate model dropdowns, and a Custom Model field overrides either. Authenticates with a short-lived JWT signed from the credential Access Key + Secret Key against the Kling AI REST API.

Category: Utility
Tool Name: kling_ai
Version: 1

Appearance: Icon: lucide-Video | Color: #6366F1

Node Type

Action (Binary) — handles file/binary data operations

Input / Output

Direction	Port(s)
Input	Input
Output	Output, Error

Credentials

This tool requires Kling AI API credentials. See the Credentials Guide for setup instructions.

Resources

Resource	Value
Image	image
Video	video
Virtual Try-On	tryOn
Account	account

Operations

Which operations are offered depends on the resource you select.

Operation	Value	Description
Create	create	Generate images from a text prompt
Get	get	Get image generation task by ID
List	list	List all image generation tasks
Text to Video	text2video	Generate video from text prompt
Image to Video	image2video	Generate video from an image
Lip Sync	lipSync	Lip sync text to a video or image
Multi-Image to Video	multiImage2video	Generate video from multiple reference images
Extend Video	videoExtend	Extend/continue an existing Kling video by ID
Video Effects	effects	Apply a single-image or dual-character video effect
Get	get	Get video task by ID
List	list	List video tasks
Create	create	Create virtual try-on
Get	get	Get try-on task by ID
List	list	List try-on tasks
Get Resource Packages	getResourcePackages	Get account resource packages and costs

The first three rows belong to Image, the next eight to Video, the next three to Virtual Try-On, and the last to Account.

Parameters

List (on Image and on Virtual Try-On) takes no parameters of its own — see All Operations. Get Resource Packages also takes none: it reports the last 30 days of resource-package usage, and an upstream item can narrow that window by carrying startTime and endTime values in Unix milliseconds.

Image: Create

Parameter	Type	Required	Default	Description
Prompt	string	No	—	Text prompt. Required by Kling for text-to-image/video and multi-image-to-video; optional for image-to-video and video extension. Supports expressions.
Negative Prompt	string	No	—	Things to exclude from the generated result. Supports expressions.
Model (image)	options	No	kling-v2-1	Kling image model. O1 / v3 Omni / v2.1 / latest v1 support text-to-image AND image-to-image (add a Reference Image below). Use Custom Model for anything else.
				Options: kling-v1-5 (Kling v1.5, latest v1), kling-v2-1 (Kling v2.1, latest v2), kling-image-o1 (Kling Image O1), kling-v3-omni (Kling v3 Omni)
Custom Model (overrides dropdown)	string	No	—	Optional. Exact model_name to send to Kling. When non-empty this OVERRIDES the Model dropdown above.
Number of Images	number	No	1	Number of images to generate (1-9).
Aspect Ratio	options	No	1:1	The aspect ratio of the generated output.
				Options: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3
Reference Image Source (image-to-image)	options	No	none	Optional reference image for image-to-image. Leave “None” for plain text-to-image. (Supported by v2.1 / Image O1 / v3 Omni / latest v1.)
				Options: none (plain text-to-image), url, base64, binary
Reference Image URL	string	No	—	URL of the reference image for image-to-image. (shown when Reference Image Source (image-to-image) is url)
Reference Image Base64	string	No	—	Base64-encoded reference image for image-to-image. (shown when Reference Image Source (image-to-image) is base64)
Reference Image Binary Field	string	No	data	Binary property holding the reference image for image-to-image. (shown when Reference Image Source (image-to-image) is binary)
Wait for Completion	boolean	No	false	Whether to poll until the task completes before returning. (Generation operations only — get/list/account return immediately.)
Polling Interval (ms)	number	No	5000	Interval in milliseconds between polling attempts. (shown when Wait for Completion is true)
Max Wait (seconds)	number	No	300	Maximum time in seconds to wait for task completion. (shown when Wait for Completion is true)
Download Result	boolean	No	false	Whether to download the result media as binary data when the task completes. (shown when Wait for Completion is true)
Output Binary Field	string	No	data	Name of the binary property to store the downloaded result. (shown when Download Result is true)

Image: Get

Parameter	Type	Required	Default	Description
Task ID	string	Yes	—	The ID of the task to retrieve. Supports expressions.

Video: Text to Video

Parameter	Type	Required	Default	Description
Prompt	string	No	—	Text prompt. Required by Kling for text-to-image/video and multi-image-to-video; optional for image-to-video and video extension. Supports expressions.
Negative Prompt	string	No	—	Things to exclude from the generated result. Supports expressions.
Model (video)	options	No	kling-v2-1	Kling video model. For anything else, use the Custom Model field below.
				Options: kling-v1-6 (Kling v1.6, latest v1), kling-v2-1 (Kling v2.1, latest v2), kling-video-o1 (Kling Video O1), kling-v3-omni (Kling v3 Omni)
Custom Model (overrides dropdown)	string	No	—	Optional. Exact model_name to send to Kling. When non-empty this OVERRIDES the Model dropdown above.
CFG Scale	number	No	0.5	How strongly generation follows the prompt (0–1). Higher = more adherence, lower = more creative freedom. Default 0.5.
Aspect Ratio	options	No	1:1	The aspect ratio of the generated output.
				Options: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3
Duration	options	No	5	Duration of the generated video in seconds.
				Options: 5 (5 seconds), 10 (10 seconds)
Mode	options	No	std	Generation mode — Standard is faster, Professional is higher quality.
				Options: std (Standard), pro (Professional)
Wait for Completion	boolean	No	false	Whether to poll until the task completes before returning. (Generation operations only — get/list/account return immediately.)
Polling Interval (ms)	number	No	5000	Interval in milliseconds between polling attempts. (shown when Wait for Completion is true)
Max Wait (seconds)	number	No	300	Maximum time in seconds to wait for task completion. (shown when Wait for Completion is true)
Download Result	boolean	No	false	Whether to download the result media as binary data when the task completes. (shown when Wait for Completion is true)
Output Binary Field	string	No	data	Name of the binary property to store the downloaded result. (shown when Download Result is true)

Video: Image to Video

Parameter	Type	Required	Default	Description
Prompt	string	No	—	Text prompt. Required by Kling for text-to-image/video and multi-image-to-video; optional for image-to-video and video extension. Supports expressions.
Negative Prompt	string	No	—	Things to exclude from the generated result. Supports expressions.
Model (video)	options	No	kling-v2-1	Kling video model. For anything else, use the Custom Model field below.
				Options: kling-v1-6 (Kling v1.6, latest v1), kling-v2-1 (Kling v2.1, latest v2), kling-video-o1 (Kling Video O1), kling-v3-omni (Kling v3 Omni)
Custom Model (overrides dropdown)	string	No	—	Optional. Exact model_name to send to Kling. When non-empty this OVERRIDES the Model dropdown above.
CFG Scale	number	No	0.5	How strongly generation follows the prompt (0–1). Higher = more adherence, lower = more creative freedom. Default 0.5.
Image Source	options	No	url	How to provide images for this operation (start frame for image-to-video; source image for an effect). The secondary images (end frame / effect 2nd image) use this same setting.
				Options: url, base64, binary
Image URL (start frame)	string	No	—	URL of the input image — the start/first frame for image-to-video. (shown when Image Source is url)
Image Base64 (start frame)	string	No	—	Base64-encoded input image (start frame). (shown when Image Source is base64)
Input Binary Field (start frame)	string	No	data	Binary property holding the input image (start frame). (shown when Image Source is binary)
End Frame Image URL (optional)	string	No	—	Optional end/tail frame URL. When set, Kling interpolates from the start frame to this end frame (image_tail). (shown when Image Source is url)
End Frame Image Base64 (optional)	string	No	—	Optional end/tail frame, base64-encoded (image_tail). (shown when Image Source is base64)
End Frame Binary Field (optional)	string	No	—	Optional binary property holding the end/tail frame (image_tail). Leave blank for none. (shown when Image Source is binary)
Duration	options	No	5	Duration of the generated video in seconds.
				Options: 5 (5 seconds), 10 (10 seconds)
Mode	options	No	std	Generation mode — Standard is faster, Professional is higher quality.
				Options: std (Standard), pro (Professional)
Wait for Completion	boolean	No	false	Whether to poll until the task completes before returning. (Generation operations only — get/list/account return immediately.)
Polling Interval (ms)	number	No	5000	Interval in milliseconds between polling attempts. (shown when Wait for Completion is true)
Max Wait (seconds)	number	No	300	Maximum time in seconds to wait for task completion. (shown when Wait for Completion is true)
Download Result	boolean	No	false	Whether to download the result media as binary data when the task completes. (shown when Wait for Completion is true)
Output Binary Field	string	No	data	Name of the binary property to store the downloaded result. (shown when Download Result is true)

Video: Lip Sync

Parameter	Type	Required	Default	Description
Lip Sync Video Source	options	No	videoId	Lip sync runs on a VIDEO: a previously generated Kling video (by ID) or an uploaded video URL.
				Options: videoId (a Kling video ID), videoUrl (an uploaded video URL)
Lip Sync Video ID	string	No	—	ID of a previously generated Kling video to lip-sync (5s/10s, kept ≤30 days). (shown when Lip Sync Video Source is videoId)
Lip Sync Video URL	string	No	—	URL of a video to lip-sync (.mp4/.mov, 2–10s, 720p/1080p, ≤100MB). (shown when Lip Sync Video Source is videoUrl)
Lip Sync Mode	options	No	text2video	Drive lip movement from generated speech (TTS) or from an audio file.
				Options: text2video (Text to Speech), audio2video (Audio File)
Lip Sync Text	string	No	—	Text to speak (TTS mode). Kling limit ~120 characters. (shown when Lip Sync Mode is text2video)
Voice ID	string	No	—	Kling TTS voice_id (e.g. “reader_en_m-v1” for English, “genshin_vindi2” for Chinese). (shown when Lip Sync Mode is text2video)
Voice Language	options	No	en	TTS voice language. (shown when Lip Sync Mode is text2video)
				Options: en (English), zh (Chinese)
Voice Speed	number	No	1	TTS speaking speed (0.8–2.0). (shown when Lip Sync Mode is text2video)
Audio URL	string	No	—	URL of an audio file to drive lip movement (.mp3/.wav/.m4a/.aac, ≤5MB). (shown when Lip Sync Mode is audio2video)
Wait for Completion	boolean	No	false	Whether to poll until the task completes before returning. (Generation operations only — get/list/account return immediately.)
Polling Interval (ms)	number	No	5000	Interval in milliseconds between polling attempts. (shown when Wait for Completion is true)
Max Wait (seconds)	number	No	300	Maximum time in seconds to wait for task completion. (shown when Wait for Completion is true)
Download Result	boolean	No	false	Whether to download the result media as binary data when the task completes. (shown when Wait for Completion is true)
Output Binary Field	string	No	data	Name of the binary property to store the downloaded result. (shown when Download Result is true)

Video: Multi-Image to Video

Parameter	Type	Required	Default	Description
Prompt	string	No	—	Text prompt. Required by Kling for text-to-image/video and multi-image-to-video; optional for image-to-video and video extension. Supports expressions.
Negative Prompt	string	No	—	Things to exclude from the generated result. Supports expressions.
Model (video)	options	No	kling-v2-1	Kling video model. For anything else, use the Custom Model field below.
				Options: kling-v1-6 (Kling v1.6, latest v1), kling-v2-1 (Kling v2.1, latest v2), kling-video-o1 (Kling Video O1), kling-v3-omni (Kling v3 Omni)
Custom Model (overrides dropdown)	string	No	—	Optional. Exact model_name to send to Kling. When non-empty this OVERRIDES the Model dropdown above.
CFG Scale	number	No	0.5	How strongly generation follows the prompt (0–1). Higher = more adherence, lower = more creative freedom. Default 0.5.
Image List	string	Yes	—	Reference images for multi-image-to-video — a JSON array of image URLs/base64 (e.g. [“https://…”,“https://…”]) or a comma-separated list. Up to ~4 images.
Duration	options	No	5	Duration of the generated video in seconds.
				Options: 5 (5 seconds), 10 (10 seconds)
Mode	options	No	std	Generation mode — Standard is faster, Professional is higher quality.
				Options: std (Standard), pro (Professional)
Aspect Ratio	options	No	1:1	The aspect ratio of the generated output.
				Options: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3
Wait for Completion	boolean	No	false	Whether to poll until the task completes before returning. (Generation operations only — get/list/account return immediately.)
Polling Interval (ms)	number	No	5000	Interval in milliseconds between polling attempts. (shown when Wait for Completion is true)
Max Wait (seconds)	number	No	300	Maximum time in seconds to wait for task completion. (shown when Wait for Completion is true)
Download Result	boolean	No	false	Whether to download the result media as binary data when the task completes. (shown when Wait for Completion is true)
Output Binary Field	string	No	data	Name of the binary property to store the downloaded result. (shown when Download Result is true)

Video: Extend Video

Parameter	Type	Required	Default	Description
Prompt	string	No	—	Text prompt. Required by Kling for text-to-image/video and multi-image-to-video; optional for image-to-video and video extension. Supports expressions.
Negative Prompt	string	No	—	Things to exclude from the generated result. Supports expressions.
CFG Scale	number	No	0.5	How strongly generation follows the prompt (0–1). Higher = more adherence, lower = more creative freedom. Default 0.5.
Source Video ID	string	Yes	—	The ID of a previously generated (succeeded) Kling video to extend/continue.
Wait for Completion	boolean	No	false	Whether to poll until the task completes before returning. (Generation operations only — get/list/account return immediately.)
Polling Interval (ms)	number	No	5000	Interval in milliseconds between polling attempts. (shown when Wait for Completion is true)
Max Wait (seconds)	number	No	300	Maximum time in seconds to wait for task completion. (shown when Wait for Completion is true)
Download Result	boolean	No	false	Whether to download the result media as binary data when the task completes. (shown when Wait for Completion is true)
Output Binary Field	string	No	data	Name of the binary property to store the downloaded result. (shown when Download Result is true)

Video: Video Effects

Parameter	Type	Required	Default	Description
Model (video)	options	No	kling-v2-1	Kling video model. For anything else, use the Custom Model field below.
				Options: kling-v1-6 (Kling v1.6, latest v1), kling-v2-1 (Kling v2.1, latest v2), kling-video-o1 (Kling Video O1), kling-v3-omni (Kling v3 Omni)
Custom Model (overrides dropdown)	string	No	—	Optional. Exact model_name to send to Kling. When non-empty this OVERRIDES the Model dropdown above.
Image Source	options	No	url	How to provide images for this operation (start frame for image-to-video; source image for an effect). The secondary images (end frame / effect 2nd image) use this same setting.
				Options: url, base64, binary
Image URL (start frame)	string	No	—	URL of the input image — the start/first frame for image-to-video. (shown when Image Source is url)
Image Base64 (start frame)	string	No	—	Base64-encoded input image (start frame). (shown when Image Source is base64)
Input Binary Field (start frame)	string	No	data	Binary property holding the input image (start frame). (shown when Image Source is binary)
Effect	options	No	squish	The Kling effect. Single-image effects use the Image above; dual-character effects (hug/kiss/heart_gesture) also need the Second Image.
				Options: squish (single image), expansion (single image), fuzzyfuzzy (single image), hug (two portraits), kiss (two portraits), heart_gesture (two portraits)
Second Image URL (dual-character effects)	string	No	—	Second portrait image URL for dual-character effects. The first image comes from the Image above. (shown when Image Source is url)
Second Image Base64 (dual-character effects)	string	No	—	Second portrait image (base64) for dual-character effects. (shown when Image Source is base64)
Second Image Binary Field (dual-character effects)	string	No	—	Binary property holding the second portrait image for dual-character effects. (shown when Image Source is binary)
Duration	options	No	5	Duration of the generated video in seconds.
				Options: 5 (5 seconds), 10 (10 seconds)
Mode	options	No	std	Generation mode — Standard is faster, Professional is higher quality.
				Options: std (Standard), pro (Professional)
Wait for Completion	boolean	No	false	Whether to poll until the task completes before returning. (Generation operations only — get/list/account return immediately.)
Polling Interval (ms)	number	No	5000	Interval in milliseconds between polling attempts. (shown when Wait for Completion is true)
Max Wait (seconds)	number	No	300	Maximum time in seconds to wait for task completion. (shown when Wait for Completion is true)
Download Result	boolean	No	false	Whether to download the result media as binary data when the task completes. (shown when Wait for Completion is true)
Output Binary Field	string	No	data	Name of the binary property to store the downloaded result. (shown when Download Result is true)

Video: Get

Parameter	Type	Required	Default	Description
Task ID	string	Yes	—	The ID of the task to retrieve. Supports expressions.
Video Type	options	No	text2video	The type of video task to query.
				Options: text2video (Text to Video), image2video (Image to Video), lip-sync (Lip Sync)

Video: List

Parameter	Type	Required	Default	Description
Video Type	options	No	text2video	The type of video task to query.
				Options: text2video (Text to Video), image2video (Image to Video), lip-sync (Lip Sync)

Virtual Try-On: Create

Parameter	Type	Required	Default	Description
Model Image Source	options	No	url	How to provide the model image for try-on.
				Options: url, binary
Model Image URL	string	No	—	URL of the model image. (shown when Model Image Source is url)
Model Image Binary Field	string	No	modelImage	Name of the binary property containing the model image. (shown when Model Image Source is binary)
Cloth Image Source	options	No	url	How to provide the clothing image for try-on.
				Options: url, binary
Cloth Image URL	string	No	—	URL of the clothing image. (shown when Cloth Image Source is url)
Cloth Image Binary Field	string	No	clothImage	Name of the binary property containing the clothing image. (shown when Cloth Image Source is binary)
Wait for Completion	boolean	No	false	Whether to poll until the task completes before returning. (Generation operations only — get/list/account return immediately.)
Polling Interval (ms)	number	No	5000	Interval in milliseconds between polling attempts. (shown when Wait for Completion is true)
Max Wait (seconds)	number	No	300	Maximum time in seconds to wait for task completion. (shown when Wait for Completion is true)
Download Result	boolean	No	false	Whether to download the result media as binary data when the task completes. (shown when Wait for Completion is true)
Output Binary Field	string	No	data	Name of the binary property to store the downloaded result. (shown when Download Result is true)

A virtual try-on request is built from the two images alone — there is no prompt or aspect ratio to set.

Virtual Try-On: Get

Parameter	Type	Required	Default	Description
Task ID	string	Yes	—	The ID of the task to retrieve. Supports expressions.

All Operations

Parameter	Type	Required	Default	Description
Max Concurrency	number	No	3	Maximum number of items to process concurrently.

Output Data

One output item per input item. The Kling task payload is merged into the item JSON at the top level, so its fields sit alongside the fields the item already carried, and binary data on the input item is forwarded.

{
  "task_id": "abc123",
  "task_status": "succeed",
  "task_result": {
    "videos": [
      { "url": "https://…/generated-video.mp4" }
    ]
  }
}

• task_id identifies the submitted job. With Wait for Completion off, generation operations return as soon as the task is accepted, so this is the only useful field — hand it to the matching Get operation later.
• task_status reports where the job stands. With Wait for Completion on, the node polls until the status is succeed and then returns; a task that reports failure, or that is still unfinished after Max Wait, is raised as an item error instead.
• task_result appears on a finished task. It carries images for image generation and try-on and videos for the video operations; each entry has a url.
• Get and List return the same task payload for the task or tasks you asked about, without polling.

With Download Result on, the first result URL is fetched and stored as binary data on the property named by Output Binary Field, as result.mp4, result.png or result.jpg depending on the media type. Binary data already on the item is kept alongside it.

Reference the result downstream by expression, e.g. {{ $json.task_id }} or {{ $json.task_result.videos[0].url }}.

Usage Examples

• Generate an AI image from a text prompt
• Restyle an image with Kling image-to-image
• Create a video from a text description
• Convert an image to a video with Kling AI
• Build a video from several reference images
• Extend an existing Kling video
• Apply a Kling video effect to a photo
• Lip sync speech onto an existing video
• Virtual try-on: overlay clothing on a model image
• Get the status of a Kling AI generation task
• List all image generation tasks

Example Configuration

Generate an image from text and wait for the result, downloading it as binary data:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "image",
    "operation": "create",
    "prompt": "A photorealistic mountain landscape at golden hour, 4K",
    "negativePrompt": "blurry, low quality, oversaturated",
    "imageModel": "kling-v2-1",
    "n": 1,
    "aspectRatio": "16:9",
    "waitForCompletion": true,
    "pollingInterval": 3000,
    "maxWaitSeconds": 120,
    "downloadResult": true,
    "outputBinaryPropertyName": "generatedImage"
  }
}

Generate images and return immediately with the task ID:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "image",
    "operation": "create",
    "prompt": "A minimalist abstract painting with cool blue tones",
    "imageModel": "kling-v1-5",
    "n": 3,
    "aspectRatio": "1:1",
    "waitForCompletion": false
  }
}

Image-to-image — restyle a reference image coming from an upstream node:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "image",
    "operation": "create",
    "prompt": "Same product on a marble surface, soft studio lighting",
    "imageModel": "kling-image-o1",
    "imageGenSource": "binary",
    "imageGenBinary": "data",
    "waitForCompletion": true,
    "downloadResult": true,
    "outputBinaryPropertyName": "restyled"
  }
}

Get an image task by ID:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "image",
    "operation": "get",
    "taskId": "{{ $json.task_id }}"
  }
}

List all image tasks:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "image",
    "operation": "list"
  }
}

Text-to-video generation:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "text2video",
    "prompt": "A cinematic drone shot flying over a dense rainforest at dawn",
    "negativePrompt": "shaky camera, pixelated, watermark",
    "model": "kling-v2-1",
    "aspectRatio": "16:9",
    "duration": "10",
    "mode": "pro",
    "waitForCompletion": true,
    "pollingInterval": 5000,
    "maxWaitSeconds": 300,
    "downloadResult": false
  }
}

Image-to-video from a URL:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "image2video",
    "model": "kling-v2-1",
    "imageSource": "url",
    "imageUrl": "https://example.com/input-photo.jpg",
    "duration": "5",
    "mode": "std",
    "waitForCompletion": true,
    "pollingInterval": 4000,
    "maxWaitSeconds": 180,
    "downloadResult": true,
    "outputBinaryPropertyName": "outputVideo"
  }
}

Image-to-video from binary data, interpolating to an end frame:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "image2video",
    "model": "kling-v1-6",
    "imageSource": "binary",
    "imageBinaryPropertyName": "uploadedImage",
    "imageTailBinary": "finalFrame",
    "duration": "10",
    "mode": "pro",
    "waitForCompletion": false
  }
}

Multi-image-to-video from several reference images:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "multiImage2video",
    "prompt": "The product rotates slowly on a pedestal",
    "model": "kling-v2-1",
    "imageList": "[\"https://example.com/front.jpg\",\"https://example.com/side.jpg\"]",
    "duration": "5",
    "mode": "std",
    "waitForCompletion": true,
    "downloadResult": true,
    "outputBinaryPropertyName": "productVideo"
  }
}

Extend an existing Kling video:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "videoExtend",
    "videoId": "{{ $json.videoId }}",
    "prompt": "The camera keeps pushing forward through the doorway",
    "cfgScale": 0.5,
    "waitForCompletion": true,
    "maxWaitSeconds": 600
  }
}

Apply a single-image effect:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "effects",
    "effectScene": "squish",
    "model": "kling-v1-6",
    "imageSource": "url",
    "imageUrl": "https://example.com/toy.jpg",
    "duration": "5",
    "mode": "std",
    "waitForCompletion": true,
    "downloadResult": true,
    "outputBinaryPropertyName": "effectVideo"
  }
}

Apply a dual-character effect, which needs a second portrait from the same source type:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "effects",
    "effectScene": "hug",
    "model": "kling-v1-6",
    "imageSource": "url",
    "imageUrl": "https://example.com/person-a.jpg",
    "effectSecondImageUrl": "https://example.com/person-b.jpg",
    "duration": "5",
    "mode": "std"
  }
}

Lip sync a previously generated Kling video with text-to-speech:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "lipSync",
    "lipSyncVideoSource": "videoId",
    "lipSyncVideoId": "{{ $json.videoId }}",
    "lipSyncMode": "text2video",
    "lipSyncText": "Thank you for watching our product demonstration.",
    "lipSyncVoiceId": "reader_en_m-v1",
    "lipSyncVoiceLanguage": "en",
    "lipSyncVoiceSpeed": 1,
    "waitForCompletion": true,
    "pollingInterval": 4000,
    "maxWaitSeconds": 360,
    "downloadResult": true,
    "outputBinaryPropertyName": "lipSyncVideo"
  }
}

Lip sync an uploaded video from an audio file:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "lipSync",
    "lipSyncVideoSource": "videoUrl",
    "lipSyncVideoUrl": "https://example.com/speaker.mp4",
    "lipSyncMode": "audio2video",
    "lipSyncAudioUrl": "https://example.com/voiceover.mp3",
    "waitForCompletion": false
  }
}

Get a video task by ID:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "get",
    "taskId": "{{ $json.task_id }}",
    "videoType": "text2video"
  }
}

List lip sync video tasks:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "video",
    "operation": "list",
    "videoType": "lip-sync"
  }
}

Virtual try-on with both images supplied as URLs:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "tryOn",
    "operation": "create",
    "tryOnModelImageSource": "url",
    "tryOnModelImageUrl": "https://example.com/model-person.jpg",
    "tryOnClothImageSource": "url",
    "tryOnClothImageUrl": "https://example.com/shirt.jpg",
    "waitForCompletion": true,
    "pollingInterval": 4000,
    "maxWaitSeconds": 180,
    "downloadResult": true,
    "outputBinaryPropertyName": "tryOnResult"
  }
}

Virtual try-on with the model image from a URL and the garment from binary data:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "tryOn",
    "operation": "create",
    "tryOnModelImageSource": "url",
    "tryOnModelImageUrl": "https://example.com/model.jpg",
    "tryOnClothImageSource": "binary",
    "tryOnClothBinaryProperty": "clothingImageData",
    "waitForCompletion": false
  }
}

Get a try-on task by ID:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "tryOn",
    "operation": "get",
    "taskId": "{{ $json.task_id }}"
  }
}

Get account resource packages — the last 30 days by default, narrowed by startTime and endTime (Unix milliseconds) when the incoming item carries them:

{
  "type": "kling_ai",
  "parameters": {
    "resource": "account",
    "operation": "getResourcePackages"
  }
}

Error Handling

Mode	Behavior
stop	Halts workflow on first error
continue	Skips failed items, passes successful ones through
errorPort	Routes failed items to Error output port

Tips

Kling AI generates images from text prompts or a reference image, converts text, one image or several into video, extends or applies effects to existing videos, overlays clothing onto model photos via virtual try-on, and lip-syncs speech onto an existing video using the Kling AI REST API. Use it when a workflow requires AI-generated visual media, product garment visualization, or lip-synced video production. Every generation is asynchronous: the node returns a task_id immediately unless Wait for Completion is on, and can download the finished media as binary data on the main output.

Key Rules Summary

• Every generation is asynchronous. Leave Wait for Completion off to submit and move on, then read the result later with the matching Get operation; turn it on to hold the item until the media is ready. Waiting stops after Max Wait (seconds), which defaults to 300 — raise it for long clips.
• Downloading requires waiting. Download Result only appears once Wait for Completion is on, and Output Binary Field only once Download Result is on.
• One source setting drives every image field in an operation. Image Source picks URL, base64 or binary for the start frame, the end frame and the second effect portrait together — provide those images in the same form.
• Custom Model wins. Anything you type into Custom Model (overrides dropdown) replaces the model chosen in the dropdown, which is how you reach a Kling model that is not listed.
• Duration values are strings. Use 5 or 10, and note that only Standard and Professional modes exist.
• Dual-character effects need two portraits. hug, kiss and heart_gesture fail without the Second Image; the other effects use the single image.
• Lip sync runs on a video, not a still image — either a Kling video ID from an earlier generation, or the URL of a video you host.