Data IO extension contract reference
The data IO (input/output) extension enables you to add data reading and writing steps to a custom Workflow Builder workflow.
You can include this extension when you register an extension app. If you register the extension app using a form, select the Data Input Output extension in the form-based UI, as shown in this figure:

If you register the extension app using an app manifest file, set the extensions.template value in your app manifest to DataIO.Version6.DataInputOutput.
This extension has five actions. The action names you specify when you register an extension app depend on whether you register the app using a form or using an app manifest file. The app manifest file values listed here appear in the actions.template property.
For details about what each action does and when it is invoked, see Data IO extension overview.
You provide the details required for action execution when you register an extension app. See Update extension and action settings for information about defining actions using the form-based registration process. See Action schema for information about action properties in the app manifest file.
You can implement data input as a standalone capability without also implementing data output. To do this when registering an extension app using a form, select Data Read as the Data Input Output Type, as shown here:

To register an extension app that reads but does not write data in an app manifest file, implement only the DataIO.Version6.SearchRecords, DataIO.Version6.GetTypeNames, and DataIO.Version6.GetTypeDefinitions actions.
Note: You cannot implement data output as a standalone capability, because it uses the data input actions to find the records to update. If your extension must write data to your system of record, you must select Data Read & Write in the form-based UI, or implement all five data IO actions in the app manifest.
Action contracts
Get Type Names (DataIO.Version6.GetTypeNames)
This action calls the external system of record to retrieve the names of the supported types. These types represent the data objects (or database table names) in your system of record. In Concerto nomenclature, these are referred to as concepts.
Note: The names of the types returned by this action will be used as input for the Get Type Definitions (DataIO.Version6.GetTypeDefinitions) action.
Get Type Names is the label for this action that appears when you register an extension app using a form. If you register your extension app using an app manifest file, the actions.template property value for this action is DataIO.Version6.GetTypeNames.
HTTP method
This action initiates a POST request to the external API service.
Request
The input is an empty JSON object. There is no data specified, because this action returns all types supported by the external system of record.
Response
Get Type Definitions (DataIO.Version6.GetTypeDefinitions)
Type definitions are based on the Concerto data modeling language. In Concerto nomenclature, types are referred to as concepts. Given a list of types, the external system must return a list of Concerto declarations representing the properties of each type. Your app proxy should initially fetch the schemas associated with these types and then convert them to Concerto’s metamodel JSON format
. This enables Docusign to interpret your chosen system’s types consistently.
You can use Concerto’s CLI or JavaScript API to convert your models.
Get Type Definitions is the label for this action that appears when you register an extension app using a form. If you register your extension app using an app manifest file, the actions.template property value for this action is DataIO.Version6.GetTypeDefinitions.
Note: When creating a RelationshipProperty, ensure that the type that is being referred to is identified.
The following examples reflect an external system of record with two tables: Dog and Human.
The Dog type is defined as:
The Breed enum has the following values:
The Human type is defined as:
HTTP method
This action initiates a POST request to the external API service.
Request
Response
GetTypeDefinitionError
Create Record (DataIO.Version6.CreateRecord)
This action creates an object in the external system of record. The type (model) is specified by typeName, and the properties are specified in the data object.
Create Record is the label for this action that appears when you register an extension app using a form. If you register your extension app using an app manifest file, the actions.template property value for this action is DataIO.Version6.CreateRecord.
HTTP method
This action initiates a POST request to the external API service.
Request
*Not always included in the JSON request. Your API proxy must handle the case in which this property is not present.
Response
Patch Record (DataIO.Version6.PatchRecord)
This action updates an existing record. Only the attributes to be updated are included in the data object.
Patch Record is the label for this action that appears when you register an extension app using a form. If you register your extension app using an app manifest file, the actions.template property value for this action is DataIO.Version6.PatchRecord.
HTTP method
This action initiates a POST request to the external API service.
Request
Response
Search Records (DataIO.Version6.SearchRecords)
This action searches for one or more records in the external system of record. It uses a custom query filtering language. Your app proxy must convert the provided query into one supported by your external system of record. Then apply the query and return the results.
Workflows use the Search Records action in Read from steps and Writeback to steps. For a Read from step, the action returns records that can be used in subsequent workflow steps, such as a Collect Data with Web Forms step that displays values from the system of record in a web form. For a Writeback to step that updates a record, a Search Records action identifies the record to be updated. For each step type, the number of records that Workflow Builder supports in the Search Records response is as follows:
Docusign recommends that extension app developers provide documentation guiding workflow process builders to use only search queries that retrieve the supported number of records.
Note: When configuring a Workflow Builder Read from step, process builders can use the preview feature to see which records the search will return. See Configure a Read from Salesforce Step for details about previewing query results.
Search Records is the label for this action that appears when you register an extension app using a form. If you register your extension app using an app manifest file, the actions.template property value for this action is DataIO.Version6.SearchRecords.
HTTP method
This action initiates a POST request to the external API service.
Example requests
Example request 1
Example request 2
Example request 3
This example request shows a query of dog objects where the breed attribute is Beagle.
Example responses
Example response 1
Example response 2
Example response 3
Multi-level querying
Multi-level querying enables retrieving data beyond the top level of an object model by accessing deeply nested properties. Unlike basic queries that return only direct attributes, multi-level queries allow for structured traversal into objects, lists, and references to extract targeted information. This approach is essential when working with complex data models where objects contain deeply nested properties or maintain relationships with other entities.
For more information about relationships, see Define related objects in a data model.
To support multi-level querying, a structured format is required to define how data should be accessed at different depths. Queries must specify whether to return full objects, individual properties, or entire sets of related attributes. One effective method for structuring these queries is through a standardized format that outlines the selection path.
This dataset represents a structured object containing personal details, addresses, and associated property information, which will be used to demonstrate how multi-level queries can retrieve specific attributes, nested properties, and relational values.
Person Table
Address Table
House Defects Table
Causes Table
This table details the supported query structures and provides concrete examples of how multi-level querying can be applied. Note that support for certain features may vary depending on the application, as some systems may not support all querying capabilities.
Error handling
If a data IO action encounters an error, it should return an HTTP error code 400, 404, or 500 along with a message describing the error.
Example error responses:
Note: A 404 error should not be returned if a Search Records request results in no matching records. Instead, return an empty records array.
This error structure applies to all actions. Developers are responsible for providing a code and message for each error.