Skip to main content

Transform Nodes

Transform nodes modify the data flowing through your workflow. They don't produce data — they reshape, filter, or route what's already there.

Map Data​

Remap fields from one format to another.

Use cases:

  • Rename fields to match your database columns
  • Extract nested values into flat fields
  • Combine multiple fields into one

Example:

Input: { "invoice_num": "INV-001", "amt": 500 }
Map: invoice_num → invoice_number, amt → total_amount
Output: { "invoice_number": "INV-001", "total_amount": 500 }

Filter​

Remove rows that don't meet a condition.

Use cases:

  • Only process invoices above $100
  • Skip documents from specific vendors
  • Filter out incomplete records

Configuration: Set a field, operator (equals, greater than, contains, etc.), and value.

Join​

Merge two data streams by matching on a common field.

Use cases:

  • Combine customer data from one source with orders from another
  • Merge extracted data with existing database records

Configuration: Select the join key field from each input stream.

Condition (If/Else)​

Branch the workflow based on a condition.

Use cases:

  • Route invoices > $10,000 to a manager approval step
  • Send different email templates based on document type
  • Process differently based on extraction confidence

Configuration: Set a field, operator, and value. The "true" branch gets data that matches; the "false" branch gets the rest.

Switch​

Like Condition but with multiple branches (more than two).

Use cases:

  • Route documents to different processors based on type (invoice, contract, receipt)
  • Send to different databases based on category

Configuration: Define cases with field values and connect each case to a different branch.

Merge​

Combine multiple input streams into one output.

Use cases:

  • After a Condition or Switch, merge the branches back together
  • Combine results from parallel processing steps
  • Aggregate data from multiple sources

Configuration: Connect multiple inputs; outputs a combined stream.