Nodes

Every workflow is made of nodes. It starts with a Trigger, and the rest fall into three families.

​

Anatomy of a node

Most generator nodes share the same parts:

• Prompt field — what you want, in words. Connected inputs appear here as labeled references like @Sketch.
• Model selector — which AI model runs. See choosing a model.
• Generation settings — options such as aspect ratio, number of results, and resolution, depending on the model.
• Run button — generate from this node.
• Output area — your results. When a node produces more than one, a navigation bar lets you step through them.
• Status — whether the node is idle, running, done, or needs attention.

A hover toolbar gives quick actions on a node’s result—annotate or inpaint, download, enhance the prompt, remove a background, and more.

​

Properties panel

When you select a node, a Properties panel opens at the bottom-right of the canvas. It exposes extra options and toggles—including settings specific to a generation model—that aren’t shown on the node itself. It’s also where you disable (freeze) a node.

​

Items and iteration

An input node can hold one item or many (several images, several lines of text). When it holds many, you decide whether each item generates its own result or they all combine into one. That choice is what lets you produce variations at scale—see variations.

​

Disabling a node

You can disable any node to freeze its output, from its Properties panel. A disabled node is skipped on the next run—no model call, no credits—but its last result stays available to everything downstream. Re-enable it when you want it to generate again.

​

Next