Skip to main content
A connector can pass the provider’s response through as-is, or map it onto an output schema you define so every provider in a category returns the same shape. That mapping is where most of the work goes, and where most of the mistakes are.

Mapping to your schema

Two step functions do the work, both documented in full on Step Functions:
Use version: "2" for both map_fields and typecast.

Give the agent your target schema

Your target schema is the thing you’d otherwise restate for every provider you map, so write it down where the agent can read it:
With that in place, the agent maps a new provider’s employee response to your schema without being asked, including the enum normalization that is easy to get subtly wrong by hand.

Checklist

Before publishing a connector with a defined output schema, verify:
  • Target schema documented before building
  • All required fields identified
  • Field types specified (string, number, enum, datetime_string)
  • Enum values defined with mappings
  • fieldConfigs map all schema fields
  • targetFieldKey uses your schema names, not the provider’s
  • Nested paths verified against the actual response
  • Enum mappings handle every provider value plus a default case
  • cursor.enabled: true for list actions
  • dataKey path verified with --debug
  • nextKey path verified with --debug
  • result.next returns the cursor for the next page
  • Tested against the first page, next page, last page, and empty results
  • map_fields step with version: "2"
  • typecast step with version: "2"
  • Correct dataSource references between steps

Build Workflow

The build and debug loop this fits into.

Step Functions

Full reference for map_fields, typecast, and the rest.

Expression Syntax

JSONPath and JEXL for the expressions in your mappings.