Search: In the top search bar, type Azure OpenAI and select it from the services list.
Azure Open AI
Initiate: Click the + Create button.
Basics Tab: Fill in the following:
Subscription: Choose your active Azure subscription.
Resource Group: Select an existing one or click "Create new."
Region: Choose a region (e.g., East US, Sweden Central) based on model availability.
Name: Give your service a unique name.
Pricing Tier: Select Standard S0.
Networking: Select your preferred network security (Standard is "All networks").
Tags:You can set up tags for consolidated billing now, or skip this step if they aren't needed.
Review + Create: Click Review + Create, then click Create once validation passes.
2.2 Deploy Your Model
Once your resource is "Successfully Deployed," follow these steps to activate a specific model (like GPT-4o).
Go to Resource: Click the Go to resource button from your deployment screen.
Launch Foundry: Under the Overview or Explore and deploy section, click on Go to Azure OpenAI Foundry (formerly AI Studio).
Deployments: In the left-hand menu, select Deployments, then click Deploy model > Deploy base model.
Select Model: Choose a Chat Completion model (e.g., gpt-4o).
Configure:
Deployment Name: Give it a name (you will need this for your code).
TPM (Tokens Per Minute): Adjust the slider based on your required capacity.
NOTE: For higher TPM limits, click Request Quota and complete the provided form to submit your application.
Finalize: Click Deploy.
2.3 Collect Your Credentials
To connect your application to this model, you need three specific pieces of information. Navigate to the Resource Management > Keys and Endpoint section in the Azure Portal:
Credential
Description
Example
Endpoint
The base URL for your resource
https://my-ai-service.openai.azure.com/
API Key
Either Key 1 or Key 2 (keep this secret!)
7a1b2c3d4e5f6g7h8i9j...
Deployment Name
The specific name you gave in Phase 2
gpt-4o-production
Security Tip: Never hard-code your API keys directly into your front-end code. Use environment variables or a Key Vault to keep them secure.
2.4 Least Privilege Permission for the User to Deploy the Model
Follow the link below to see a summary of the Azure OpenAI Roles and Figure 1 shows an example of these permissions:
Login to the Lepide Web Console and the Home page will be displayed:
Lepide Web Console Home Screen
Click the Settings icon Lepide Admin Console (top right of the screen) to display the Admin Console
Configure Database
From the Categories menu on the left-hand side, select Lepide AI (you may need to scroll down the menu to see the Lepide AI Option)
Click on Configure Database
Database Credentials for Lepide AI
Enter the database credentials:
Lepide AI Screen
Enter the SQL Server name
There are two authentication options:
Windows Authentication: Add the domain account which has at least dbowner rights on the SQL server or
SQL Server Authentication: Select this mode if SQL Server is installed on a remote machine or the local machine. We recommend that this option is selected.
Provide the username and password of a SQL User. This user must have sufficient rights to create a new database.
Click Test Connection to test the connection between the solution and the selected SQL Server.
It will either display an error if connection fails or a message confirming successful connection.
There are two Database options:
Create Database: Enter a database name in the Create Database field to create a new database or
Select Database: Select an existing database created by Lepide or another application
Select the Data Retention Period: The options are 15, 30, 60 or 90 days
Select the Time Out settings for Connection and Query
Can't find what you're looking for? Access our previous PDF-based library.
4 min read
️ Updated Jul 09, 2026
