Nodes

Most canvases store shapes. Lines Nodes store meaning.

A Node is a structured piece of your product model. It can represent a user, task, screen, workflow step, decision, rule, object, feature, or any other concept your team needs to understand. Connect Nodes to show how those concepts depend on one another, then use the model as the source of truth for design and implementation.

Describe more than a label

Each Node can carry the context behind an idea, including its name, description, semantic type, inputs, outputs, state, permissions, rules, and relationships. That structure keeps important decisions attached to the work instead of scattering them across documents, diagrams, and chat threads.

Use Nodes to answer questions such as:

• Who is involved?

• What are they trying to accomplish?

• What information enters and leaves each step?

• Which rules or permissions apply?

• What can happen next?

• Where does a decision affect the experience?

Build a model people can follow

Arrange Nodes spatially on the Model canvas, nest them to create hierarchy, and connect them to describe a flow. Sections and sticky notes help organize the model without turning presentation elements into product logic.

The same hierarchy is available in the sidebar, so you can move between a high-level system map and the details of an individual Node without losing your place.

Keep meaning connected to design

Nodes live in Model Mode. Visual Views live in Design Mode. Lines keeps the two concerns distinct and lets you bind them during the Package and Ship stage.

That separation matters: one product concept can appear in more than one interface, and one interface can express several parts of the product model. The model remains understandable even as the visual design changes.

Give agents usable context

A flat mockup shows what something looks like. A connected Node model also explains what it means, what it depends on, and what constraints apply. This gives implementation agents a clearer starting point and gives people an inspectable record of the intent behind the output.

Nodes are stored in the local Lines Markdown file alongside their relationships and supporting context. You can inspect the file yourself and explicitly package the model when you are ready to hand work off.

Learn how to work with Nodes in the documentation — /docs/nodes.