Nodes and Connections

Understand the major node types, how they behave, and how data moves across the canvas.

Nodes and Connections

Nodes are the building blocks of a flow. Each one represents a step in your workflow.

Main node types

Generative nodes

These call AI models and usually consume credits when run:

Image Generation
Video Generation
Audio
Music Generation

These nodes usually include:

model selection
prompt input
model-specific settings
run state and output preview

Input and passthrough nodes

These help feed data into the flow:

Upload
Text Input
Output

Upload gives the flow source media.
Text Input supplies text to other nodes.
Output marks the final end of a branch.

Composition and editing nodes

These help combine or modify media:

Composer
Video Transform
Audio Transform
Image Transform
Overlay
Video Editor

These are useful when you already have generated or uploaded content and want to refine it.

Connection types

LooksCraft uses typed handles:

image
video
audio
text

Connections only work when the source and target handle types match.

How required inputs work

Some models allow optional inputs. Others require a specific source, such as:

an image input
a video source
an audio file
a text prompt

If a node cannot run, check:

missing required connections
unsupported input type
incompatible model choice for your current input

Model-driven node behavior

For generative nodes, the chosen model controls:

what settings appear
which inputs are allowed or required
estimated runtime
pricing behavior

This means two nodes of the same type can expose different controls depending on the model.

Special behavior examples

Some video utilities depend on an upstream generated task or source video.
Some image models support reference images, while others are text-only.
Some audio/music nodes behave more like utilities than pure generation.

How to design clean flows

Use this pattern:

source nodes on the left
generation nodes in the middle
transform/editor nodes after generation
output nodes on the right

That keeps your flow easier to read and debug.

Good starter combinations

Text Input -> Image Generation -> Output
Upload -> Video Generation -> Output
Upload -> Audio -> Composer -> Output
Upload -> Image Transform -> Video Generation -> Output