Video node

The video node generates AI video from a text prompt together with image and video inputs. The available input methods, video length, ratio, resolution, and sound settings depend on the model you choose.

Inputs and options depend on the model

When you choose an AI model in the video node, only the input slots and generation options that model supports are available.

Choose the model first

Changing the model also changes which start frame, end frame, reference image, and video input are available, along with the length, ratio, resolution, and sound options.

How to use

Choose a model, set the inputs and options that model supports, then generate the video.

  1. Add a video node.

    Add a video node in Canvas.

  2. Choose an AI model.

    Choosing a model shows that model's input slots and the options at the bottom.

  3. Connect the prompt and any media you need.

    Connect a start frame, end frame, reference image, or video input, within what the model supports.

  4. Set the generation options.

    Set the options shown, such as length, ratio, resolution, and sound.

  5. Generate the video.

    The result appears in the center of the video node.

  6. Choose what to do next.

    Once the video is finished, an output point appears on the right of the node. Select it and choose the next task from the menu below.

Menu itemNode added
Add to an editA Composition node with this video connected
Edit with AIA Gemini Omni Flash video node with this video connected to its video input. Describe the change in the prompt and generate.
Post to YouTubeA YouTube upload node. Not available if one is already connected.

Input methods

Depending on the model, some of the inputs below are available.

InputDescription
PromptEnter the scene, motion, mood, camera movement, and so on as text.
Start frameSpecify the image to use as the video's first scene.
End frameSpecify the image to use as the video's last scene.
Reference imageEnter an image for the generated video to reference, such as a person, product, or style.
Video inputUsed with models that can take an existing video as input data.

Generation options

Choose the model and the video generation options at the bottom of the node. The choices shown differ by model.

Video length

Set the generated length within the range the model supports.

Ratio

On supported models you choose the ratio directly. Some models follow the start frame's ratio.

Resolution

Choose from 360p, 480p, 720p, 768p, 1080p, 2K, and 4K, depending on what the model supports.

Sound

Depending on the model, sound is supported or generated automatically from the prompt.

Support by model

The input methods and generation options for each model in the Klipse video node.

ModelStartEndReferenceVideo inputLengthRatioResolutionSound
Seedance 2.0 MiniSupportedSupportedUp to 9Supported4–15s9:16, 16:9, 1:1, 21:9, 4:3, 3:4480p, 720pSupported
Seedance 2.0 FastSupportedSupportedUp to 9Supported4–15s9:16, 16:9, 1:1, 21:9, 4:3, 3:4480p, 720pSupported
Seedance 2.0SupportedSupportedUp to 9Supported4–15s9:16, 16:9, 1:1, 21:9, 4:3, 3:4480p, 720p, 1080p, 4KSupported
Seedance 2.5SupportedSupportedUp to 30Supported4–30s9:16, 16:9, 1:1, 21:9, 4:3, 3:4480p, 720p, 1080pSupported
Kling 2.6SupportedSupportedNot supportedNot supported3–15sFollows start frame720p, 1080pSupported
Kling 3.0SupportedSupportedNot supportedNot supported3–15sFollows start frame720p, 1080p, 4KSupported
Kling OmniSupportedSupportedUp to 7Supported3–10s9:16, 16:9, 1:1720p, 1080p, 4KSupported
Kling 4.0 Coming soonSupportedSupportedUp to 10Supported3–30s9:16, 16:9, 1:1, 21:9720p, 1080p, 4KSupported
MiniMax H3SupportedSupportedUp to 9Supported4–15s9:16, 16:9, 1:1, 21:9, 4:3, 3:4768p, 2KAutomatic, prompt-based
HappyHorse 1.1 Coming soonSupportedNot supportedUp to 9Not supported3–15s9:16, 16:9, 1:1, 21:9, 9:21, 4:3, 3:4, 4:5, 5:4480p, 720p, 1080pAutomatic, prompt-based
Gemini Omni FlashSupportedNot supportedUp to 6Supported3–10s9:16, 16:9360p, 720p, 1080p, 4KAutomatic, prompt-based
Veo 3.1 ProSupportedSupportedUp to 3Not supported4, 6, 8s9:16, 16:9720p, 1080pAutomatic, prompt-based
Veo 3.1 FastSupportedSupportedUp to 3Not supported4, 6, 8s9:16, 16:9720p, 1080pAutomatic, prompt-based
Veo 3.1 LiteSupportedSupportedNot supportedNot supported4, 6, 8s9:16, 16:9720p, 1080pAutomatic, prompt-based

Seedance 2.5

Seedance 2.5 generates the same way the 2.0 series does, but a single clip can run twice as long and it accepts far more reference images.

Supports the 9:16, 16:9, 1:1, 21:9, 4:3, and 3:4 ratios, and accepts up to 30 reference images. It does not support 4K. When a start frame is connected, the ratio follows the start frame.

  • 4–30s
  • up to 1080p
0:00
−0:00
Choosing 2.5 raises the length slider's ceiling to 30 seconds. Longer clips and higher resolutions cost more credits.

Seedance 2.0 Series

The Seedance family generates video around the prompt, with the length, ratio, and resolution chosen directly.

Seedance 2.0 Mini / Fast

Supports the 9:16, 16:9, 1:1, 21:9, 4:3, and 3:4 ratios.

  • 4–15s
  • 480p / 720p

Seedance 2.0

Supports the same ratios, with a choice of 480p, 720p, 1080p, and 4K.

  • 4–15s
  • up to 4K
0:00
−0:00
The credit cost differs by model and updates whenever you change the length, ratio, or resolution.

Kling 2.6 / Kling 3.0

Specify a start frame and an end frame to control the video's first and last scenes.

Kling 2.6

  • Start
  • End
  • 3–15s
  • 720p / 1080p
  • Sound supported

Kling 3.0

  • Start
  • End
  • 3–15s
  • 720p / 1080p / 4K
  • Sound supported
0:00
−0:00
Switching from Kling 3.0 to Kling 2.6 lets you compare the length and resolution choices of the two models.

The ratio follows the start frame

Kling 2.6 and Kling 3.0 have no separate ratio choice; the video ratio is determined by the ratio of the start frame image.

Kling Omni

You can use reference images and a video input together, and the video input is used as a reference or for AI Edit. With a video input connected, reference images are limited to 4, and start and end frames are not available.

Supported inputs

  • Start frame
  • End frame
  • Reference image
  • Video input

Generation options

  • 3–10s
  • 9:16 / 16:9 / 1:1
  • 720p / 1080p / 4K
  • Sound supported
0:00
−0:00
Kling Omni takes both a reference image and a video input, so its input row differs from the other models.

Kling 4.0

Coming soon

You can set the flow of scenes with start and end frames and up to 10 keyframes, and add a range of references to make video of up to 30 seconds.

A keyframe is an input that sets a key scene of the video, which makes it different from a reference image the video draws a character or style from. As references you can use up to 10 images, up to 5 videos (30 seconds in total), up to 7 elements such as a person or product (up to 3 of them video-based), and a voice, with up to 15 in all.

Supports the 9:16, 16:9, 1:1, and 21:9 ratios and 720p, 1080p, and 4K resolution, and sound comes out as two-channel stereo. A video input can be used as a reference or for editing an existing video — changing expressions, motion, camera, background, style, and more.

  • Start frame
  • End frame
  • Up to 10 keyframes
  • Reference image
  • Video input
  • Voice reference
  • 3–30s
  • 720p / 1080p / 4K
  • Stereo audio

Kling 4.0 Flash

Coming soon

Makes 3–20 second videos at 720p quickly from text or an image, which suits generating drafts again and again.

It can also generate with references.

  • Image-based generation
  • Reference
  • 3–20s
  • 720p

MiniMax H3

You can create video from a prompt alone, or add start and end frames or reference images and video inputs. Sound is generated automatically from the prompt.

An omni-modal model that reads text, images, and video in one pass. A character can come from a reference image, and the scene and movement from a video input.

Supports the 9:16, 16:9, 1:1, 21:9, 4:3, and 3:4 ratios and 768p and 2K resolution. You can connect up to 9 reference images and up to 3 video inputs as references. With reference images or video inputs connected, start and end frames are not available. When a start frame is connected, the ratio follows the start frame. Sound has no on/off setting: it is always generated from the prompt together with the video, and comes out as native stereo.

  • Start frame
  • End frame
  • Reference image
  • Video input
  • 4–15s
  • 768p / 2K
  • Automatic audio generation
0:00
−0:00
Choosing MiniMax H3 sets the length range to 4–15 seconds and offers 768p and 2K resolutions.

HappyHorse 1.1

Coming soon

You can create video from a prompt alone, or add a start frame or reference images. Sound is generated automatically from the prompt.

Generates video from a prompt alone, or from a start frame or reference images, keeping a character, product, outfit, or brand element in the same shape and style through the video.

Supports the 9:16, 16:9, 1:1, 21:9, 9:21, 4:3, 3:4, 4:5, and 5:4 ratios. You can connect one start frame or up to 9 reference images, but not both at once. End frames and video input are not supported. When a start frame is connected, the ratio follows the start frame. Sound has no on/off setting: it is always generated from the prompt, with lip sync matched automatically.

  • Start frame
  • Reference image
  • 3–15s
  • 480p / 720p / 1080p
  • Automatic audio generation

Gemini Omni Flash

You can create 3–10 second video from a start frame and reference images. Video input is used for AI Edit only.

Supported inputs

  • Start frame
  • Reference image
  • Video input

Generation options

Sound is generated automatically from the prompt. When you use AI Edit with a video input, the result follows the input video's length and ratio. Choose Edit with AI from a finished video's output point to add a video node set to this model.

  • 3–10s
  • 9:16 / 16:9
  • 360p / 720p / 1080p / 4K
  • Automatic audio generation
0:00
−0:00
Choosing Gemini Omni Flash lowers the length slider's maximum to 10 seconds, and resolution goes up to 4K.

Veo 3.1 Series

The Veo 3.1 family lets you specify start and end frames, and sound is generated automatically from the prompt. Pro and Fast can use reference images instead of a start frame, in which case the length is fixed at 8 seconds.

Veo 3.1 Pro / Fast

  • Start
  • End
  • Reference
  • 4 / 6 / 8s
  • 9:16 / 16:9
  • 720p / 1080p

Veo 3.1 Lite

Length, ratio, and resolution are the same as Pro/Fast, but reference images are not supported.

  • Start
  • End
  • 4 / 6 / 8s
  • 9:16 / 16:9
  • 720p / 1080p
0:00
−0:00
Switching models leaves the length, ratio, and resolution choices unchanged.

Tips

Choose the model by input method first.

Deciding first whether you need an end frame, a reference image, or a video input makes choosing a model easier.

Check the final ratio and resolution first.

Check that the model supports the ratio your destination needs — 9:16 for Shorts/Reels, 16:9 for standard video.

For Kling 2.6 / 3.0, check the start frame ratio.

Instead of a separate ratio choice, these two models let the start frame's ratio decide the video ratio.

With automatic audio models, describe the sound in the prompt too.

MiniMax H3, HappyHorse 1.1, Gemini Omni Flash, and the Veo 3.1 family generate sound automatically from the prompt.

Troubleshooting

ProblemWhat to check
The input slot you want is not shownCheck the per-model support table to see whether the model you chose supports that input method.
The ratio menu is not shownCheck whether the model follows the start frame's ratio, as Kling 2.6 / 3.0 do.
The ratio shows Auto and cannot be changedSeedance 2.5, MiniMax H3, and HappyHorse 1.1 follow the start frame's ratio once a start frame is connected. When you use AI Edit with Gemini Omni Flash, the ratio follows the input video.
You cannot choose the resolution you wantCheck the resolution range each model supports.
You cannot connect a reference imageCheck whether the model supports reference images, as the Seedance series, Kling Omni, Kling 4.0, MiniMax H3, HappyHorse 1.1, Gemini Omni Flash, and Veo 3.1 Pro/Fast do. On the Seedance series, MiniMax H3, HappyHorse 1.1, and Veo 3.1 Pro/Fast, reference images cannot be used together with a start frame.

Frequently asked questions

Can I use a start frame and an end frame on every model?

No. Support differs by model. Check the support table above.

Which models support reference images?

The Seedance series, Kling Omni, Kling 4.0, MiniMax H3, HappyHorse 1.1, Gemini Omni Flash, Veo 3.1 Pro, and Veo 3.1 Fast support reference images. The Seedance 2.0 series, MiniMax H3, and HappyHorse 1.1 accept up to 9, Kling 4.0 up to 10, and Seedance 2.5 up to 30.

Can I input a video directly?

The Seedance series, Kling Omni, Kling 4.0, MiniMax H3, and Gemini Omni Flash accept a video input. The Seedance series and MiniMax H3 use it as a reference, Kling Omni as a reference or for AI Edit, and Gemini Omni Flash for AI Edit only. Kling 4.0 uses a video input as a reference or to edit an existing video.

Which models support 4K?

Seedance 2.0, Kling 3.0, Kling Omni, Kling 4.0, and Gemini Omni Flash support 4K.