Microsoft AI Foundry
Configure Microsoft AI Foundry as an AI provider for chat and embedding capabilities.
Microsoft AI Foundry (formerly Azure AI Studio) provides a unified endpoint for accessing multiple AI models within the Azure ecosystem, supporting both Chat and Embedding capabilities.
Installation
Install-Package Umbraco.AI.MicrosoftFoundryOr via .NET CLI:
dotnet add package Umbraco.AI.MicrosoftFoundryConnection Settings
The provider supports two authentication methods: Entra ID (recommended) and API Key (legacy).
Entra ID Authentication (Recommended)
Endpoint
Yes
Your AI Foundry endpoint URL
Project Name
No
AI Foundry project name (enables deployed model listing)
Tenant ID
No
Microsoft Entra ID tenant ID
Client ID
No
Service principal application (client) ID
Client Secret
No
Service principal secret
Entra ID supports two modes:
Service Principal - Provide Tenant ID, Client ID, and Client Secret for explicit credentials
Managed Identity - Provide only Tenant ID (or leave all Microsoft Entra ID fields empty) to use
DefaultAzureCredential. This supports managed identities, Azure CLI, and other automatic credential sources
API Key Authentication (Legacy)
Endpoint
Yes
Your AI Foundry endpoint URL
API Key
Yes
Your AI Foundry API key
Advanced Settings
Use Responses API
No
When enabled, uses the OpenAI Responses API instead of Chat Completions. Only available in certain Azure regions.
Getting Your Credentials
For Entra ID
Sign in to Azure AI Studio
Open your AI project
Copy the Project Name and Endpoint from the project overview
Create a service principal in Microsoft Entra ID > App registrations
Grant the Azure AI Developer role to the service principal on your AI Hub resource
For API Key
Sign in to Azure AI Studio
Open your AI project
Navigate to Deployments
Select your deployment
Copy the Target URI (endpoint) and Key

Related
Last updated
Was this helpful?