How to Upload Files to OpenAI Storage Using the OpenAI .NET SDK

In this OpenAI blog post, We’ll show you how to upload multiple files to OpenAI’s storage using C#. We’ll walk through a simple console application that reads files from a folder and uploads them to OpenAI.

If you’re building applications with the OpenAI API, there are scenarios where you need to upload files — for example, when using Assistants, fine-tuning models, or providing external documents. OpenAI provides secure storage for your uploaded files, and with the OpenAI .NET SDK, the process is straightforward.

Prerequisites

Before you start, ensure you have:

• A valid OpenAI API key
• The latest OpenAI .NET SDK installed (OpenAI NuGet package)
• Basic familiarity with C# and asynchronous programming

You can install the OpenAI .NET SDK by running:

dotnet add package OpenAI

Step 1: Set Up the OpenAI Clients

First, we create an OpenAIClient using your API key, which should be securely stored in an environment variable called OPENAI_API_KEY. This ensures sensitive credentials are not hardcoded.

OpenAIClient openAIClient = new(Environment.GetEnvironmentVariable("OPENAI_API_KEY"));
OpenAIFileClient fileClient = openAIClient.GetOpenAIFileClient();
AssistantClient assistantClient = openAIClient.GetAssistantClient();

• OpenAIClient is the main entry point.
• OpenAIFileClient handles all file upload operations.
• AssistantClient is created if you also want to work with the Assistants API.

Step 2: Prepare Files for Upload

Next, specify the folder containing the files you want to upload. This example assumes a folder named Files exists at the application root:

string folderPath = "Files";
string[] filePaths = Directory.GetFiles(folderPath);

This line fetches all file paths from the folder, so you can loop over them.

Step 3: Upload Files to OpenAI

Using a foreach loop, the application reads each file as a FileStream and uploads it using UploadFileAsync:

foreach (string filePath in filePaths)
{
    using FileStream fileStream = File.OpenRead(filePath);

    OpenAIFile uploadedFile = await fileClient.UploadFileAsync(
        fileStream,
        Path.GetFileName(filePath),
        FileUploadPurpose.Assistants);

    Console.WriteLine($"Uploaded File: {uploadedFile.Filename}, ID: {uploadedFile.Id}");
}

Key points:

• Files are uploaded with the purpose Assistants, which is used when files are associated with OpenAI Assistants workflows.
• After uploading, the filename and its OpenAI-generated ID are printed for reference.

Step 4: List Uploaded Files

Finally, you might want to verify the files you uploaded by listing all files currently stored in your OpenAI account:

try
{
    var result = await fileClient.GetFilesAsync();
    var files = result.Value;

    foreach (var file in files)
    {
        Console.WriteLine($"File ID: {file.Id}, Filename: {file.Filename}, Size: {file.SizeInBytes} bytes");
    }
}
catch (Exception ex)
{
    Console.WriteLine($"Error: {ex.Message}");
}

This retrieves and prints a list of all uploaded files, including their IDs, filenames, and sizes.

Summary

Uploading files to OpenAI using the .NET SDK is simple and efficient. This approach can be integrated into larger projects where you want to automate the management of document libraries, fine-tuning datasets, or file attachments for AI assistants.

Tip: Always remember to manage uploaded files properly. Unused files can accumulate over time and may impact your storage limits and costs.

If you’re interested in extending this, you could add features like:

• Deleting old or unused files
• Categorizing uploads by project
• Monitoring upload status in real time

Happy coding!

For OpenAI and .NET consulting please contact us