Use the Call REST API tool to add OpenAPI-based REST operations to AI Agents and Copilots. By importing a YAML or JSON schema, you can define multiple API operations at once and allow your AI-powered experiences to call approved external APIs using structured operations.
The Call REST API tool helps Agents and Copilots:
Access external business systems
Discover available API operations from an OpenAPI schema
Use structured REST operations instead of unmanaged HTTP requests
Reuse multiple API endpoints from a single schema configuration
Before you start
Before configuring the Call REST API tool, make sure:
You have already created an available API connection
The connection type is either:
HTTP API (Generic)
OAuth 2.0 API
You have a valid OpenAPI schema in YAML or JSON format
You have permission to edit the AI Agent or Copilot
Add the Call REST API tool
Open the AI Agent or Copilot you want to configure.
Go to the Tools tab.
Click Add a tool.
Select Call REST API.
Click Next.
The configuration window for the Call REST API tool opens.
Select a connection and create the tool
In the Call REST API setup window:
Select a connection.
Review the connection type.
Enter a tool name.
Click Add or Add and configure.
Field descriptions
Field | Description |
Select connection | Select an available API connection. |
Connection type | Displays the selected connection type. |
Name | The display name of the tool inside the AI Agent or Copilot. |
The available connection scope is consistent with the Call HTTP request tool.
Supported connection types include:
HTTP API (Generic)
OAuth 2.0 API
Configure tool details
After creating the tool, the full configuration page opens.
In the Details section, you can:
Update the tool name
Add or edit the tool description
Review the selected connection
View the connection type
Check the connection status
If the selected connection uses OAuth 2.0 authentication, authorization information is also displayed.
Define the API schema
The Schema section is used to define the available REST API operations.
Supported schema formats
The Call REST API tool supports:
YAML
JSON
The schema must follow the OpenAPI specification format.
An OpenAPI schema is a standard API definition document that describes:
Available API endpoints
Request methods
Parameters
Request and response structures
Operation summaries
Add the schema
Go to the Schema section.
Paste the OpenAPI schema into the schema editor.
Yeeflow automatically parses the schema.
Detected operations appear in the Available operations section.
When the schema is valid, Yeeflow automatically discovers the available operations that the Agent or Copilot can use.
YAML example
openapi: 3.1.0 info: title: Sample API version: 1.0.0 servers: - url: https://api.example.com/v1 paths: /users: get: summary: Get all users parameters: - name: page in: query schema: type: integer responses: '200': description: Success
JSON example
{ "openapi": "3.1.0", "info": { "title": "Sample API", "version": "1.0.0" }, "servers": [ { "url": "https://api.example.com/v1" } ] }Review available operations
The Available operations section displays all operations parsed from the schema.
The table includes:
Column | Description |
Name | Operation name |
Method | HTTP method |
Path | API endpoint path |
Configuration | Customize operation settings |
Use this section to verify that the schema has been parsed correctly.
Customize operation settings
Click Customize in the Configuration column to edit the operation settings.
You can customize:
Operation name
Operation description
By default, the operation description uses the summary value from the schema definition.
Customizing operation names and descriptions can help the AI Agent or Copilot better understand when to use each operation.
Configure completion behavior
The Completion section controls what happens after the tool finishes running.
You can configure:
After running behavior
Output variables available to the Agent and other tools
Default outputs include:
Output | Type |
Data | Object |
Status code | Number |
These outputs can be referenced by downstream tools or workflow steps.
Preview and test the tool
Use the Preview & test run panel to test the tool configuration before publishing.
You can:
Enter input variables
Run a test execution
Verify that operations are selected correctly
Review returned API results
Testing is recommended before enabling the tool for production use.
Save and publish
After configuration is complete:
Click Save.
Test the AI Agent or Copilot.
Click Publish when ready.
Published changes become available to end users after deployment.
Notes
Use clear and action-oriented operation names and descriptions.
Only include operations that the Agent or Copilot should be allowed to use.
Test all operations before publishing.
Complete OAuth authorization if required by the selected connection.
Use OpenAPI schemas to manage related API operations together.
For simple single-endpoint scenarios, the Call HTTP request tool may be more suitable than Call REST API.