Mapping or Transforming Data

1. Direct Mapping

Click the data icon beside a destination field and choose a value from Data select, or embed it in text:

{{ 2.email }}
Order {{ 2.order_id }} is ready for collection.

Paths follow each Step’s stored output. Native app actions return values directly, while HTTP Request nests its response under ret. Webhook nests the payload under body unless Body only is enabled. Skip extra Steps when the source already fits.

2. Single-Value Pipes

Open the field’s data icon and switch to Data transformation, or write pipes inside the braces. Operations run left to right without changing the upstream Step’s output:

{{ 2.email | trim | lower }}
{{ 3.subtotal | times:1.21 | number_format:2:sep_comma_dot }}
{{ 4.company | default:"Unknown" }}

Run math before formatting a number as text. Note that default replaces missing, null, empty, or whitespace-only values while keeping 0 and false. See Data Transformation.

In-Field List Pipes

When only one field needs a quick list summary, chain list pipes in place:

{{ 3.results | sort:ascending:name | pluck:email | join:comma_space }}

This sorts records by name, extracts each email, and joins them with commas. Use a dedicated List Step instead when multiple Steps reuse the list or you need to inspect items in execution history.

3. Parse JSON

Use Parse JSON when an upstream field holds raw JSON text like {"customer":{"email":"[email protected]"},"total":125}. Parse JSON returns the parsed structure directly at Step 4:

{{ 4.customer.email }}
{{ 4.total }}
{{ 4.items.0.sku }}

Both dot (4.items.0.sku) and bracket (4.items[0].sku) indexing work. Skip Parse JSON when the data picker already exposes the fields.

4. List Steps

Use a dedicated list Step when the reshaped collection drives routing, loops, or multiple downstream Steps:

5. AI Data Transform

Use AI Data Transform only when layout varies across runs, such as emails or receipts. Write up to three extraction instructions:

Extract the purchase order number from {{ 2.ret.body }}.

At Step 5, outputs are positional ({{ 5.result.0 }}, {{ 5.result.1 }}, {{ 5.result.2 }}). Reordering instructions changes downstream mappings. Prefer deterministic pipes or Parse JSON when the format is fixed.

6. Code Editor

Use Code editor when multiple inputs interact or Python is clearer than a long chain of Steps. Keep references standalone and unquoted:

import json
 
orders = {{ 3.results }}
tax_rate = {{ $var.TAX_RATE }}
 
total = sum(order["amount"] for order in orders)
print(json.dumps({"total_with_tax": round(total * tax_rate, 2)}))

Code editor parses the final JSON line of stdout automatically. Map {{ 4.result.output.total_with_tax }} directly, or read raw stdout at {{ 4.result.executionOutput }} without Parse JSON.