The image node is the most common thing on the DOTV canvas: every storyboard frame in an AI video, every character portrait, every scene reference lives in an image node — one picture, plus where it came from. This section is about the parts that are its own: the shape, the toolbar, images you upload, and where images come from. Creating a node, what the card looks like and the four states are the same for every node, and they live in 2.1 and 2.4.
An image node's frame reshapes to the ratio you pick: 16:9 is landscape, 9:16 is portrait, 1:1 is square.
Element edit and Layer split, both inside Edit, have no chapter of their own: the first points at the spots to change (click / box / brush), says what each should become or where it should move, and regenerates the whole list in one pass; the second breaks the picture into a stack of transparent layers (person, props, background) you can drag / resize / flip before merging them back into one image.
The storyboard side uses this same bar — open an image's details and it sits along the top edge of the picture (see 1.4).
Publish style (the share icon on the right) sends this image's look, along with the settings it was made with, to the style library, where anyone can pick it up and apply it. For the other direction — picking a style for an image of your own — see 3.2.
Select an image node that has already generated and a toolbar floats above the picture — every entry point in this chapter on images is on it.
From left to right:
| Button | What it does | Covered in |
|---|---|---|
| Preset | Fifteen ready-made recipes in four groups — storyboard, texture, space & camera, sheets. Skin texture and the 720° panorama are both in here. | 3.6 |
| Edit | Everything that touches the picture itself: Upscale, Outpaint, Repaint, Erase, Cut out, plus Element edit and Layer split. The button shows whichever one you used last. | 3.3 / 3.4 |
| Reshoot | Expression tuning, Multi-angle, Relight — these three change how it was shot, not what is in the picture. | 3.5 |
| Split grid | Cut one multi-cell collage into separate images. | 3.5 |
| The six on the right | Annotate (circle / box / caption a spot — saved as a new image node on the right), Rotate & flip (90° turns and mirrors, saved as a new node), Crop — none of these three touch a model; then Publish style, Download, Full screen. | 3.3 |
An image brought in from your own machine is an image node too; it just has no prompt and can't "regenerate". This kind of node has its own Replace media entry at the top right of the picture: swap another image in and the node is still the same node, with every wire out of it left alone.
Besides writing a prompt yourself (see 3.2), these also drop image nodes onto the canvas: