Video Node
A Video Node holds a single video clip. You can upload a clip from your computer or generate one using an AI video model. Video Nodes are typically the final step in a workflow — they receive text, images, and other videos as input and produce the finished video output.
Adding content
Uploading a video
When a Video Node is empty, click the Upload button in the display area to select a file from your computer.
Supported formats: .mp4, .mov
Maximum file size: 500 MB
After uploading, the node retains its generation panel and both connection points. If you generate a new video using the model, it replaces the uploaded clip.
Generating with a model
Write a prompt in the generation panel, select a generation mode and model, then click Generate. The video appears in the display area once the model returns a result.
The Generate button is disabled when:
- There is no prompt and no upstream content connected
- A video is already being generated on this node
- An upstream node is currently generating ("Upstream task in progress.")
Generation modes
The Video Node offers three generation modes. The active mode determines how your upstream connections and prompt are interpreted.
Create Video
Generates a video from a text prompt, with no image or video reference required. Use this when starting from scratch or when a Text Node provides the creative direction.
This mode is selected automatically when only Text Nodes are connected upstream.
First & Last Frame
Generates a video that starts on one reference image and ends on another. Requires exactly two upstream Image Nodes — the first connected becomes the first frame, the second becomes the last frame.
Select this mode manually after connecting two Image Nodes. Any connected Video Nodes are ignored in this mode.
Omni-Reference
Generates a video using multiple reference images and/or videos simultaneously. The model draws on all connected upstream assets to produce a result that reflects their combined visual style, characters, and scenes.
This mode is selected automatically when two or more Image Nodes are connected upstream, or when any Video Node is connected upstream.
Tip: Omni-Reference is the most flexible mode — use it to maintain visual consistency across scenes by feeding character or environment images from earlier nodes.
Generation panel
| Element | Description |
|---|---|
| Reference strip | Thumbnails of all connected upstream Image and Video Nodes, numbered in connection order |
| Prompt input | Your generation instructions; placeholder: "Describe the video you want to generate." |
| Upstream text icons | Inline icon blocks from connected upstream Text Nodes |
| Mode selector | Create Video, First & Last Frame, or Omni-Reference |
| Model selector | Video model; incompatible models are greyed out based on connected inputs |
| Parameters | Resolution, aspect ratio, duration, audio options (vary by model) |
| Generate button | Shows credit cost (credits × seconds × resolution) before you confirm |
Auto mode and auto model switching
When you connect or disconnect upstream nodes, the mode switches automatically to stay compatible:
- Only Text Nodes, or one Image Node → Create Video
- Two or more Image Nodes, or any Video Node → Omni-Reference
If your current model doesn't support the active mode or the number of connected inputs, the system switches to a compatible model automatically. See Video Models for each model's limits.
Model input limits
Each video model accepts a different number of reference images and videos. If your connected inputs exceed a model's limit, that model is greyed out — switch to one with a higher limit, or remove a reference. For the full per-model breakdown, see Video Models.
Connecting to other nodes
Accepted upstream connections
| Upstream node | How its content is used |
|---|---|
| Text Node | Text becomes an inline prompt icon; up to 99 Text Nodes |
| Image Node | Image appears in the reference strip; numbered in connection order |
| Video Node | Video appears in the reference strip as a first-frame thumbnail |
Supported downstream connections
A Video Node can connect downstream to another Video Node or a Text Node. It cannot connect downstream to an Image Node — no image model currently supports video input.
Display area toolbar
| State | Toolbar options |
|---|---|
| Empty / generating / failed | Upload button only |
| Video present | Full screen, Download |
Content and deletion
When you delete a Video Node, its video is saved to your account assets — it is not lost. You can access saved videos from your assets library.
Updated 2 days ago