Editor

Nodes

Node types, properties, and how to place them on the canvas.

Nodes are the building blocks of a diagram. Each node has a type, visual style, and optional metadata.

Add a node

  1. Pick a type from the palette.
  2. Drop or click to place it on the canvas.
  3. Edit its title and properties in the inspector.

Common kinds

KindRole
Infra / componentApps, services, databases, queues, gateways, etc.
GroupFrame that contains other nodes
NoteFree text / comment on the canvas
PortLink stub / inherited connection point

Infra nodes often carry:

  • typeId (app, postgres, kafka, …)
  • Technologies (tags from a catalog)
  • Status (healthy, degraded, …)
  • Scope (internal / external)
  • Lifecycle (live / future / deprecated / removed)
  • Accent color
  • Properties (key/value facts, optionally shown on canvas)
  • Connectors (published in/out handles from a child diagram)

Properties that matter

Title — primary label on the card.

Subtitle / description — extra context; description may support markdown in the inspector.

displayDescription — short caption on the canvas when you do not want the full description.

Tags — freeform labels used with filters and perspectives.

Technologies — catalog IDs (e.g. postgres, kubernetes) for consistent icons/labels.

Properties — structured facts:

  • key / value (e.g. OS → Linux, IP → 10.0.0.1)
  • showOnCanvas — show or keep inspector-only
  • optional icon id

Child diagram — opens a nested canvas; see Nested diagrams.

Visual style

Use shape, accent, and lifecycle to encode meaning:

  • Dashed / muted styles often mean external, future, or deprecated
  • Status colors communicate health without opening the inspector
  • Keep accents consistent across a team template

Editing tips

  • Prefer stable titles (Checkout API) over temporary ones (tmp1).
  • Put environment-specific facts in properties, not in the title.
  • Use notes for decisions; use infra nodes for systems.
  • Multi-select and align when cleaning a messy layout.