ApiRequest

The API Request activity enables seamless integration with AI models to automate intelligent tasks such as content generation, text summarization, data classification, sentiment analysis, language translation, decision support, and other AI-driven insights.

It securely retrieves the API key from a configured vault at runtime, ensuring that sensitive credentials are never exposed during design or execution.

Limitation

The API key must be configured only through the Vault in Robility Manager using the Pin & Key configuration.
Currently, this activity supports integration only with Azure and its supported AI models.

Security Behavior

1. The API key is securely retrieved from the configured vault only during workflow execution. 
2. The API key is not visible in: Activity Designer UI, Workflow variables & Logs or trace outputs
3. The API key exists only in memory during execution and is cleared immediately after use.
4. The API key is never logged, displayed, or persisted in any form.

Parameters

INPUT

Tenant Name: Automatically populated when the project is connected using the Publish
Connector*: Connector options become available only after the project is connected and can be selected from the drop-down menu.
Vault Key*: Vault key options become available only after the project is connected and can be selected from the drop-down menu.
API Version*: Specifies the version of the API used for processing requests.
Deployment Name*: Specifies the name of the deployed model used to generate responses.
Endpoint*: Specifies the base URL of the API service used for processing requests.
System Prompt*: Provides content to guide the AI in a specific role, behavior, or tone, and defines how responses should be generated (e.g., technically, creatively).
  Example: Acts as a content creator and generates content in a user-friendly way to explain a software product to new users in 5 simple lines.
User Prompt*: Provides input for the AI to generate a response based on the System Prompt.

Note: All input values can be hardcoded, assigned using a String variable, or passed as a String datatype.

MISC

DisplayName : Displays the name of the activity.
SkipOnError: Specifies how the workflow behaves when an error occurs:
   True – Continues workflow execution even if an error occurs.
   False – Stops workflow execution when an error occurs.
   None (Default) – Behaves the same as False.
Version: Indicates the version of the activity.

OUTPUT

Response: Specifies the response returned after processing the request.
ResponseContent: Specifies the actual response content in JSON format, to be viewed in detail.
Note: Response and ResponseContent are available only as String variables.
Result: Returns the execution status of the activity as a Boolean value:
    True – Activity executed successfully.
    False – Activity failed due to an error.

* Represents mandatory fields required to execute the workflow.

How to Do?

Step 1: Connect with the Project  
     You need to select the project to connect.
    Once connected, you can choose the Connector and the Vault Key (a secure key to access sensitive information).

Step 2: Configure the Activity and Execute the Workflow 
    After connecting, provide all the required information the activity needs to work in each parameter.
    Then, run the workflow so the AI will response with your input and give the output as instructed, system can perform the task automatically to give the response from the AI.

Use Cases

1. Automated content creation (emails, reports, descriptions) 
2. Text summarization of large documents
3. Question answering based on given input
4. Automatic data classification and tagging
5. End-to-end workflow automation using AI