Skip to main content
Openprise Academy

Microsoft Foundry Data Service

Updated

Pro & Enterprise

Purpose

In the Microsoft Foundry Data Service, an LLM chat service and credentials need to be configured. The service is used by the AskAI task template. 

 

Prerequisites

  1. You must have an active Azure subscription with a valid payment method in order to sign-in with your Microsoft account credentials.  
  2. You will need to deploy a model in Microsoft Azure AI Foundry prior to creating a data service in Openprise. 

Detailed instructions on how to deploy a model in your Azure AI Foundry environment can be found HERE. Additional instructions regarding deploying a fine-tuned model can be found HERE.

Locate Model-specific Information From Deployed Model

You will need the following model-specific information from the deployed model you want to connect to Openprise:

  • Target URI
  • Model name
  • API Key

Select the resource group to which your model belongs. This can be changed by navigating to the selection drop-down in the top left and selecting the desired resource group. 

foundry2.1.png
  1. Select Build in the top right corner.
  2. Select Deployments on the left menu bar.
  3. Select Deployed models and then select the model name (e.g. gpt-5-mini model).
foundry2.2.png
  1. Navigate to Playground under the model name.
  2. Select Code. 
  3. Select Python from the Language dropdown. Select OpenAI SDK from the SDK dropdown. 
  4. Select an API type from the dropdown. If Responses API is listed, the model supports web search and you will be able to configure whether or not to use web search when creating a data service in Openprise. If Completions API is only available, web search will NOT be supported by the model.

foundry2.3.png

Copy the endpoint and deployment name. You will need them for when you configure your data service. NOTE: Only copy the text within the quotes NOT the quotes themselves.

Navigate to Details under the model name. Make note of the following information: 

foundry2.5.png

  1. Target URI: Copy the provided URI for authentication in Openprise. You will only enter a portion of this URL up to and including the “/openai/” segment (e.g. https://op-azureopenai.openai.azure.com/openai/).
  2. Copy the associated model API key. Store this key in a secure location.
  3. Model name: note the name of the model you are trying to connect to in Openprise (e.g. gpt-4o-mini)

Create Microsoft Foundry Data Service in Openprise

Navigate to Data > Data Services.

select data services.png

Select Add Data Service.

add data service.png

Give your Data Service a name, optional description, assign data service administrators, and labels. Select Microsoft Foundry in the Select technology section. Click Add Account Information. Select a registered account, or add a new account using your Microsoft Azure AI Foundry credentials. 

foundry1.6.png

To add a new Account, follow the steps below:

foundry1.7.png

  1. Email: Enter your Microsoft Azure AI Foundry email address.
  2. AI model: Enter the name of the AI model you want Openprise to connect to. 
  3. API Key: Paste the API Key from your Microsoft Foundry account.
  4. Use Web Search: Select "Yes" from the dropdown if your model supports web search. Select "No" if your model does not support web search. 
  5. API Endpoint: Paste the Target URI from your Microsoft Foundry account. Only enter a portion of this URL up to and including the “/openai/” segment (e.g. https://op-azureopenai.openai.azure.com/openai/). 

Click Save in the top right corner to save your data service configuration. You will now see a new Data Service tile created on the main Data Services screen in Openprise. 

foundry1.8.png

Related articles