Video generation node
Basic configuration
Purpose: Generate video based on text, image, video or audio input.
01
Add video node.
02
Different models support different reference methods and parameters. For details, please refer to03.2 Video model and common parameters。
- 01Video model: Click to switch to generate model
- 02Reference mode: select all-purpose reference or first and last frame modes
- 03Aspect ratio: set landscape, portrait or square
- 04Resolution: Set the output resolution
- 05Duration: Set the length of the generated video
- 06Generation quantity: Set how many videos to generate at one time
- 07Audio switch: turn on or off synchronized audio
03
ChooseAll-round referencemode.
- 01Reference material: connect pictures, videos or audio
- 02Prompt word: Enter the picture and action you want to generate
- 03Generate parameters: set model, scale, duration, etc.
- 04
Run: Start generating video
04
ChooseFirst and last framesmode.
- 01First frame: Connect the beginning of the video
- 02End frame: Connect the video ending screen
- DescriptionConnect the first and last frame pictures, and then select the first and last frame mode.
Video model parameter matrix
DATA TABLEVideo model parameters
| model | Material reference form | What is suitable for generating | resolution | duration | The quantity you can choose to generate at one time |
|---|---|---|---|---|---|
| MiniMax H3 | Universal reference, first and last frames | It supports all-round reference and is cost-effective, suitable for videos that require a variety of reference materials. | 720p | 4–15 seconds | 1 |
| Gemini Omni Flash | All-purpose reference, you can specify the first frame | Emphasis on physical perception and consistency, suitable for the use of picture or video reference generation. | 720p | 4–10 seconds | 1 |
| Seedance 2.5 | All-round reference, first and last frames, intelligent editing (can partially edit existing videos), super long video (can generate 180s long video at one time) | The latest official multi-modal Video generation, suitable for scenarios where you want to get a complete video quickly. | 480p、720p、1K、4K | 4–30 seconds | 1–4 |
| Seedance 2.0 | Universal reference, first and last frames | Supports picture, video and audio reference, suitable for multi-modal material creation. | 720p、1080p、4K | 4–15 seconds | 1–4 |
| Seedance 2.0 (fast version) | Universal reference, first and last frames | Fast multimodal Video generation up to 720p for quick previews. | 720p | 4–15 seconds | 1–4 |
| SeedanceLite2 | Universal reference, first and last frames | Close to the native effect of Seedance 2, supporting up to 4K, suitable for cost control. | 720p、1080p、4K | 4–15 seconds | 1–4 |
| SeedanceLite2 Fast | Universal reference, first and last frames | Fast output to 720p for low-cost, fast previews. | 720p | 4–15 seconds | 1–4 |
| Seedance 1.0 Pro Fast | Only one picture can be uploaded as the first frame | The generation speed is fast, the movement is smooth, the picture is exquisite, and it is suitable for quick preview. | 720p、1080p | 2–12 seconds | 1–4 |