Skip to content

Latest commit

 

History

History
186 lines (128 loc) · 7.1 KB

azureai-openai-integration.md

File metadata and controls

186 lines (128 loc) · 7.1 KB
title description ms.topic ms.date
.NET Aspire Azure OpenAI integration (Preview)
Learn how to use the .NET Aspire Azure OpenAI integration.
how-to
02/14/2025

.NET Aspire Azure OpenAI integration (Preview)

In this article, you learn how to use the .NET Aspire Azure OpenAI client. The Aspire.Azure.AI.OpenAI library is used to register an OpenAIClient in the dependency injection (DI) container for consuming Azure OpenAI or OpenAI functionality. It enables corresponding logging and telemetry.

For more information on using the OpenAIClient, see Quickstart: Get started generating text using Azure OpenAI Service.

Get started

To get started with the .NET Aspire Azure OpenAI integration, install the 📦 Aspire.Azure.AI.OpenAI NuGet package in the client-consuming project, i.e., the project for the application that uses the Azure OpenAI client.

dotnet add package Aspire.Azure.AI.OpenAI
<PackageReference Include="Aspire.Azure.AI.OpenAI"
                  Version="*" />

For more information, see dotnet add package or Manage package dependencies in .NET applications.

Example usage

In the :::no-loc text="Program.cs"::: file of your client-consuming project, call the extension method to register an OpenAIClient for use via the dependency injection container. The method takes a connection name parameter.

builder.AddAzureOpenAIClient("openAiConnectionName");

In the preceding code, the AddAzureOpenAIClient method adds an OpenAIClient to the DI container. The openAiConnectionName parameter is the name of the connection string in the configuration. You can then retrieve the OpenAIClient instance using dependency injection. For example, to retrieve the connection from an example service:

public class ExampleService(OpenAIClient client)
{
    // Use client...
}

App host usage

To add Azure hosting support to your xref:Aspire.Hosting.IDistributedApplicationBuilder, install the 📦 Aspire.Hosting.Azure.CognitiveServices NuGet package in the app host project.

dotnet add package Aspire.Hosting.Azure.CognitiveServices
<PackageReference Include="Aspire.Hosting.Azure.CognitiveServices"
                  Version="*" />

In your app host project, register an Azure OpenAI resource using the following methods, such as xref:Aspire.Hosting.AzureOpenAIExtensions.AddAzureOpenAI%2A:

var builder = DistributedApplication.CreateBuilder(args);

var openai = builder.ExecutionContext.IsPublishMode
    ? builder.AddAzureOpenAI("openAiConnectionName")
    : builder.AddConnectionString("openAiConnectionName");

builder.AddProject<Projects.ExampleProject>()
       .WithReference(openai);

The AddAzureAIOpenAI method will read connection information from the app host's configuration (for example, from "user secrets") under the ConnectionStrings:openAiConnectionName config key. The xref:Aspire.Hosting.ResourceBuilderExtensions.WithReference%2A method passes that connection information into a connection string named openAiConnectionName in the ExampleProject project. In the :::no-loc text="Program.cs"::: file of ExampleProject, the connection can be consumed using:

builder.AddAzureAIOpenAI("openAiConnectionName");

Configuration

The .NET Aspire Azure OpenAI integration provides multiple options to configure the connection based on the requirements and conventions of your project.

Use a connection string

When using a connection string from the ConnectionStrings configuration section, you can provide the name of the connection string when calling builder.AddAzureAIOpenAI:

builder.AddAzureAIOpenAI("openAiConnectionName");

The connection string is retrieved from the ConnectionStrings configuration section, and there are two supported formats, either the account endpoint used in conjunction with the default Azure credential or a connection string with the account key.

Account endpoint

The recommended approach is to use an Endpoint, which works with the AzureOpenAISettings.Credential property to establish a connection. If no credential is configured, the xref:Azure.Identity.DefaultAzureCredential is used.

{
  "ConnectionStrings": {
    "openAiConnectionName": "https://{account_name}.openai.azure.com/"
  }
}

For more information, see Use Azure OpenAI without keys.

Connection string

Alternatively, a custom connection string can be used.

{
  "ConnectionStrings": {
    "openAiConnectionName": "Endpoint=https://{account_name}.openai.azure.com/;Key={account_key};"
  }
}

In order to connect to the non-Azure OpenAI service, drop the Endpoint property and only set the Key property to set the API key.

Use configuration providers

The .NET Aspire Azure OpenAI integration supports xref:Microsoft.Extensions.Configuration. It loads the AzureOpenAISettings from configuration by using the Aspire:Azure:AI:OpenAI key. Example :::no-loc text="appsettings.json"::: that configures some of the options:

{
  "Aspire": {
    "Azure": {
      "AI": {
        "OpenAI": {
          "DisableTracing": false,
        }
      }
    }
  }
}

Use inline delegates

Also you can pass the Action<AzureOpenAISettings> configureSettings delegate to set up some or all the options inline, for example to disable tracing from code:

builder.AddAzureAIOpenAI(
    "openAiConnectionName",
    static settings => settings.DisableTracing = true);

You can also setup the OpenAIClientOptions using the optional Action<IAzureClientBuilder<OpenAIClient, OpenAIClientOptions>> configureClientBuilder parameter of the AddAzureAIOpenAI method. For example, to set the client ID for this client:

builder.AddAzureAIOpenAI(
    "openAiConnectionName",
    configureClientBuilder: builder => builder.ConfigureOptions(
        options => options.Diagnostics.ApplicationId = "CLIENT_ID"));

[!INCLUDE integration-observability-and-telemetry]

Logging

The .NET Aspire Azure OpenAI integration uses the following log categories:

  • Azure
  • Azure.Core
  • Azure.Identity

See also