Skip to main content
Nodes are the building blocks of the canvas. Each one holds a piece of content — a prompt, an image, a clip, a sound — and you wire them together so one node’s output becomes the next node’s input.

​
Node anatomy

Every node shares the same shape: a header with the type icon and name, the content area where the result appears, connection handles on the left (input) and right (output) edges, and a prompt bar with a Run button at the bottom. The ⋮ menu in the top-right corner holds per-node actions, and the pills around the prompt bar hold that node’s generation settings — aspect ratio, resolution, duration, and so on.
A canvas node showing its header, content, handles, and footer

An Image node, showing its header, content area, side connection handles, and prompt-and-Run footer

​
Add a node

Click the + button in the left toolbar to open the add-node menu. Generate lists the AI node types, Tools lists the image tools, and Libraries holds Upload and Asset Library for bringing your own files onto the canvas. The menu is searchable, and most entries have a shortcut so you can drop a node without opening it at all:
The add-node menu showing Generate and Libraries sections with keyboard shortcuts

The add-node menu, opened from the + button in the left toolbar

​
Connect nodes

Each node has an input handle on its left edge and an output handle on its right. Click and drag from an output handle to another node’s input handle — a curved edge appears, and the source now feeds the target.
A connection being dragged between two nodes' handles

Dragging from a Text node's output handle to an Image node's input handle

There’s a faster way when the next node doesn’t exist yet: hover a handle and a radial menu fans out with every node type that connection supports. Pick one, and Flowy creates the new node already wired up. To remove a connection, select the edge and press Delete.

​
Run a generation

Write a prompt in the node’s prompt bar and click Run. The node shows its progress, and when the generation finishes the result appears in the content area — the node is auto-named from your prompt so your canvas stays readable.

​
How outputs feed inputs

Connections are how context flows. Everything wired into a node becomes material for its next run: connect Text → Image and the text drives the image prompt; connect Image → Video and the image conditions the clip; connect Audio → Composition and the sound lands on the timeline. Each node type accepts specific inputs and produces specific outputs — the radial menu only ever offers compatible types, so you can’t wire something a node can’t use.

​
Node types

Text

Prompts, scripts, and notes that feed every other node.

Image

Generate stills up to 4K, with references and photobooth mode.

Video

AI video clips with two engines: Default and Compose.

Audio

Sound effects, music, and voiceover from a prompt.

3D

3D models from images and text, with a built-in orbit viewer.

Vector

SVG graphics — generate from text or vectorize an image.

Flam

Alpha-channel motion clips you can composite over anything.

Storyboard

Scene-by-scene plans with a story bible, fanned out to nodes.

Composition

A layered timeline inside the canvas that opens in Studio.

Utility nodes

Router, Group, Sticky Note, Brand Kit, and Model nodes.
Not sure which node to start with? Ask the AI assistant to plan and add nodes for you.