> For the complete documentation index, see [llms.txt](https://loova.gitbook.io/infinite-canvas/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://loova.gitbook.io/infinite-canvas/create-on-the-canvas/video-nodes.md).

# Video Nodes

Use Video nodes to animate a storyboard, generate a clip from references, modify existing footage, or continue a video.

<figure><img src="/files/O6qzIlHQFGgpA9GOIra8" alt="Video generation controls with a connected reference" width="90%"><figcaption></figcaption></figure>

### Choose the right generation mode

The mode selector changes with the selected model. Not every model supports every path.

| Mode                         | Use it when…                                                          | Typical input                                        |
| ---------------------------- | --------------------------------------------------------------------- | ---------------------------------------------------- |
| **Text to Video**            | You want to explore motion without a fixed starting composition       | Text prompt                                          |
| **Frames to Video**          | You want to animate a storyboard frame or move between defined frames | First frame; optional last frame when supported      |
| **Reference to Video**       | Identity, style, objects, or visual cues should guide the result      | Text, images, and supported video references         |
| **Reference Video to Video** | A source video should guide motion or visual behavior                 | Reference video plus prompt and supported references |
| **Video Edit**               | You want to describe changes to an existing video                     | Source video plus edit instruction                   |
| **Video Extend**             | You want the action to continue beyond the current ending             | Source video plus continuation prompt                |

Switching models updates the available modes, durations, resolutions, aspect ratios, and advanced settings.

<figure><img src="/files/3KOy0HSh1KXDPumSl4kB" alt="Selecting a video generation mode" width="90%"><figcaption></figcaption></figure>

### Assign first and last frames

Connect an Image node to a Video node, then select the connection label and choose:

* **First Frame**
* **Last Frame**
* **Reference**

A Video node can have only one First Frame and one Last Frame. Assigning an occupied role to another image swaps the existing role automatically.

<figure><img src="/files/GZ96m3k72MwwHsId8cLc" alt="Assigning a connected image as a video frame or reference" width="90%"><figcaption></figcaption></figure>

### Write a useful video prompt

Focus on change over time. Describe the subject motion, camera motion, environmental motion, pacing, and the details that must be preserved. If a result is unstable, simplify the number of simultaneous actions before adding more instructions.

### Video node actions

* Upload or replace a video
* Generate, edit, or extend when supported by the selected model
* Play the clip inside the node
* Reference it from downstream Text, Image, or Video generation
* Rename, download, duplicate, or delete the node