Edit

Share via

Transcribe recorded audio files with Foundry Local

Important

  • Foundry Local is available in preview. Public preview releases provide early access to features that are in active deployment.
  • Features, approaches, and processes can change or have limited capabilities, before General Availability (GA).

In this article, you learn how to use Foundry Local's native audio transcription API to convert recorded audio files into text. You create a C# console application that downloads a transcription model called Whisper, loads it, and transcribes an audio file using the Whisper model. By the end of this article, you understand how to integrate audio transcription capabilities into your local applications without requiring cloud connectivity.

Prerequisites

Samples repository

The sample in this article can be found in the Foundry Local C# SDK Samples GitHub repository.

Create project

Use Foundry Local in your C# project by following these Windows-specific or Cross-Platform (macOS/Linux/Windows) instructions:

  1. Create a new C# project and navigate into it: 
    dotnet new console -n app-name
cd app-name
  2. Open and edit the app-name.csproj file to: 
    <Project Sdk="Microsoft.NET.Sdk">

  <PropertyGroup>
    <OutputType>Exe</OutputType>
    <TargetFramework>net9.0-windows10.0.26100</TargetFramework>
    <RootNamespace>app-name</RootNamespace>
    <ImplicitUsings>enable</ImplicitUsings>
    <Nullable>enable</Nullable>
    <WindowsAppSDKSelfContained>false</WindowsAppSDKSelfContained>
    <WindowsPackageType>None</WindowsPackageType>
    <EnableCoreMrtTooling>false</EnableCoreMrtTooling>
  </PropertyGroup>

  <ItemGroup>
    <PackageReference Include="Microsoft.AI.Foundry.Local.WinML" Version="0.8.2.1" />
    <PackageReference Include="Microsoft.Extensions.Logging" Version="9.0.10" />
    <PackageReference Include="OpenAI" Version="2.5.0" />
  </ItemGroup>

</Project>
  3. Create a nuget.config file in the project root with the following content so that the packages restore correctly: 
    <?xml version="1.0" encoding="utf-8"?>
<configuration>
  <packageSources>
    <clear />
    <add key="nuget.org" value="https://api.nuget.org/v3/index.json" />
    <add key="ORT" value="https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/ORT/nuget/v3/index.json" />
  </packageSources>
  <packageSourceMapping>
    <packageSource key="nuget.org">
      <package pattern="*" />
    </packageSource>
    <packageSource key="ORT">
      <package pattern="*Foundry*" />
    </packageSource>
  </packageSourceMapping>
</configuration>

Use audio transcription API

The following example demonstrates how to use the native audio transcription API in Foundry Local. The code includes the following steps:

  1. Initializes a FoundryLocalManager instance with a Configuration.
  2. Gets a Model object from the model catalog using an alias. Note: Foundry Local selects the best variant for the model automatically based on the available hardware of the host machine.
  3. Downloads and loads the model variant.
  4. Uses the native audio transcription API to generate a response.
  5. Unloads the model.

Copy-and-paste the following code into a C# file named Program.cs:

using Microsoft.AI.Foundry.Local;
using Microsoft.Extensions.Logging;

var config = new Configuration
{
    AppName = "app-name",
    LogLevel = Microsoft.AI.Foundry.Local.LogLevel.Debug
};

using var loggerFactory = LoggerFactory.Create(builder =>
{
    builder.SetMinimumLevel(Microsoft.Extensions.Logging.LogLevel.Debug);
});
var logger = loggerFactory.CreateLogger<Program>();

// Initialize the singleton instance.
await FoundryLocalManager.CreateAsync(config, logger);
var mgr = FoundryLocalManager.Instance;

// Get the model catalog
var catalog = await mgr.GetCatalogAsync();

// Get a model using an alias and select the CPU model variant
var model = await catalog.GetModelAsync("whisper-tiny") ?? throw new System.Exception("Model not found");
var modelVariant = model.Variants.First(v => v.Info.Runtime?.DeviceType == DeviceType.CPU);
model.SelectVariant(modelVariant);

// Download the model (the method skips download if already cached)
await model.DownloadAsync(progress =>
{
    Console.Write($"\rDownloading model: {progress:F2}%");
    if (progress >= 100f)
    {
        Console.WriteLine();
    }
});

// Load the model
await model.LoadAsync();

// Get a chat client
var audioClient = await model.GetAudioClientAsync();

// get a cancellation token
CancellationToken ct = new CancellationToken();

// Get a transcription with streaming outputs
var response = audioClient.TranscribeAudioStreamingAsync("Recording.mp3", ct);
await foreach (var chunk in response)
{
    Console.Write(chunk.Text);
    Console.Out.Flush();
}
Console.WriteLine();

// Tidy up - unload the model
await model.UnloadAsync();

Note

You'll need to replace "Recording.mp3" with the path to your audio file that you want to transcribe. Foundry Local has a native audio transcription API that allows you to transcribe audio files in the following formats:

  • WAV
  • MP3
  • FLAC

Run the code using the following command:

If your architecture is x64, use the following command:

dotnet run -r:win-x64

If your architecture is arm64, use the following command:

dotnet run -r:win-arm64

Related content

  • Last updated on