Transforms — The Basics

A transform is a JSON object with a "type" field naming the operation. The engine reads the type, runs it against your attributes, and returns a single value.

Shape

1

Rules that apply to every type

• Any string anywhere in the definition may contain ${attr} placeholders — inputs, values, mapping results, defaults, rule outcomes.
• A bare string used where a value is expected is treated as a template. So "${department}" and { "type": "static", "value": "${department}" } are equivalent.
• Most types take an input. Inside a pipeline, a step with no input receives the previous step's result automatically.
• If a transform cannot produce a value — malformed JSON, an unknown type, an unparseable date, or an empty result — the attribute is simply omitted from what is sent, rather than failing the whole operation. The editor validates the JSON as you type so you catch mistakes early.
• Nesting and pipelines are limited to 20 levels deep, a guard against a self-referential definition.

The catalogue