Data and variables

Each execution path carries a payload and the values made available along that path. A node can transform its incoming payload, emit a new one, or route it onward unchanged.

Payloads and node outputs#

Payloads can be objects, arrays, strings, numbers, booleans, or null. Their shape depends on the node. Select an executed node to see the actual value before writing an expression.

Node scopes expose an input and an output. For example, {{start.output.startTime}} reads the Start output. A Variable node normally has the scope variable; additional Variable nodes use names such as variable2. Use the expression picker to select the correct node and field.

For nodes with several outputs, the port name can be part of the output structure. Refer to the generated node reference and the picker instead of guessing a path.

Named variables#

The Variable node assigns names to values. If it creates customerName, later nodes on that path can read {{customerName}}. Its output object contains the variables it just assigned.

Variables belong to an execution path. A value created on one sibling branch is not automatically available to another sibling. Connect the dependency explicitly.

Choose the right lifetime#

• Variables: intermediate values for the current execution path.
• Globals: reusable configuration available when a run begins.
• State: workflow memory that persists between runs on a particular runtime.

Use a variable for a calculated message, a Global for a configurable threshold, and State for a saved cursor. Keep credentials in connections or the wallet vault, rather than ordinary values.

Next: expressions, Globals, and persistent State.