Transform Node

Reshape, filter, map, validate, convert formats, and run expressions on workflow data.

The Transform node modifies data between steps. Open the node and click Select Transform to choose from the transformation catalog.

TransformWhat it does
AI TransformDescribe changes in plain English; AI applies them to your data
CodeRun JavaScript or Python expressions on the input
Edit FieldsRename, add, or remove fields (key-value mapping)
FilterKeep or remove items matching conditions
AggregateSum, average, count, min, or max over a list
Parse JSONParse a JSON string into an object
Format ConversionConvert between JSON, CSV, XML, YAML, etc.

Edit Fields (mapping)

Map input data to a new shape:

Output fieldValue
fullName{{lastOutput.first}} {{lastOutput.last}}
email{{lastOutput.contact.email}}
timestamp{{new Date().toISOString()}}

Each row creates a field on the output object.

Code transform

Write a JavaScript expression that receives the input and returns the transformed result:

return {
  items: lastOutput.records.filter(r => r.active),
  total: lastOutput.records.length,
};

Scope variables (lastOutput, inputs, variables) are available in the expression.

Filter

Define conditions to include or exclude items:

  • Field: status, Operator: equals, Value: active
  • Combine multiple conditions with AND/OR logic

Format conversion

Convert data between formats:

  • JSON → CSV (for spreadsheet export)
  • XML → JSON (for API processing)
  • Markdown ↔ HTML (for content pipelines)

Data category transforms

Additional transforms available in the catalog:

CategoryExamples
Date & TimeFormat dates, add/subtract time, extract parts
TextSplit, join, replace, trim, case conversion
ValidationValidate against a schema, check required fields
SanitizationStrip HTML, escape characters, remove sensitive fields
MergeCombine multiple inputs (append, merge by key)
Sort & LimitSort arrays, take first N items
CryptoHash, HMAC, encrypt/decrypt
CompressionCompress or decompress data

Chaining transforms

For complex pipelines, chain multiple Transform nodes:

HTTP Request → Transform (Parse JSON) → Transform (Edit Fields) → Action

Tips

  • Use Edit Fields for simple renaming — it's faster than writing code.
  • Use AI Transform when the transformation is hard to express as rules ("make this text more formal").
  • Inspect input/output in the Execution Data tab after each transform to verify the shape.
  • When an AI node returns JSON as a string, add a Parse JSON transform before using the fields.