Skip to content

C# / .NET API Reference

The C# package is a pure .NET HTTP client targeting .NET 8+. No FFI or native libraries required.

Installation

dotnet add package LiterLlm

Client

Constructor

using LiterLlm;

var client = new LlmClient(
    apiKey: Environment.GetEnvironmentVariable("OPENAI_API_KEY")!,
    baseUrl: "https://api.openai.com/v1",  // default
    maxRetries: 2,                          // default
    timeout: TimeSpan.FromSeconds(60)       // default
);

LlmClient implements IDisposable and IAsyncDisposable:

await using var client = new LlmClient(apiKey: "sk-...");
Parameter Type Default Description
apiKey string required API key for Authorization: Bearer
baseUrl string https://api.openai.com/v1 Provider base URL
maxRetries int 2 Retry count for 429/5xx
timeout TimeSpan? 60s Request timeout

Methods

All methods are async and accept an optional CancellationToken.

ChatAsync(request, ct)

Task<ChatCompletionResponse> ChatAsync(ChatCompletionRequest request, CancellationToken ct = default)

EmbedAsync(request, ct)

Task<EmbeddingResponse> EmbedAsync(EmbeddingRequest request, CancellationToken ct = default)

ListModelsAsync(ct)

Task<ModelsListResponse> ListModelsAsync(CancellationToken ct = default)

ImageGenerateAsync(request, ct)

Task<ImagesResponse> ImageGenerateAsync(CreateImageRequest request, CancellationToken ct = default)

SpeechAsync(request, ct)

Returns raw audio bytes.

Task<byte[]> SpeechAsync(CreateSpeechRequest request, CancellationToken ct = default)

TranscribeAsync(request, ct)

Task<TranscriptionResponse> TranscribeAsync(CreateTranscriptionRequest request, CancellationToken ct = default)

ModerateAsync(request, ct)

Task<ModerationResponse> ModerateAsync(ModerationRequest request, CancellationToken ct = default)

RerankAsync(request, ct)

Task<RerankResponse> RerankAsync(RerankRequest request, CancellationToken ct = default)

CreateFileAsync(request, ct)

Task<FileObject> CreateFileAsync(CreateFileRequest request, CancellationToken ct = default)

RetrieveFileAsync(fileId, ct)

Task<FileObject> RetrieveFileAsync(string fileId, CancellationToken ct = default)

DeleteFileAsync(fileId, ct)

Task<DeleteResponse> DeleteFileAsync(string fileId, CancellationToken ct = default)

ListFilesAsync(query?, ct)

Task<FileListResponse> ListFilesAsync(FileListQuery? query = null, CancellationToken ct = default)

FileContentAsync(fileId, ct)

Task<byte[]> FileContentAsync(string fileId, CancellationToken ct = default)

CreateBatchAsync(request, ct)

Task<BatchObject> CreateBatchAsync(CreateBatchRequest request, CancellationToken ct = default)

RetrieveBatchAsync(batchId, ct)

Task<BatchObject> RetrieveBatchAsync(string batchId, CancellationToken ct = default)

ListBatchesAsync(query?, ct)

Task<BatchListResponse> ListBatchesAsync(BatchListQuery? query = null, CancellationToken ct = default)

CancelBatchAsync(batchId, ct)

Task<BatchObject> CancelBatchAsync(string batchId, CancellationToken ct = default)

CreateResponseAsync(request, ct)

Task<ResponseObject> CreateResponseAsync(CreateResponseRequest request, CancellationToken ct = default)

RetrieveResponseAsync(responseId, ct)

Task<ResponseObject> RetrieveResponseAsync(string responseId, CancellationToken ct = default)

CancelResponseAsync(responseId, ct)

Task<ResponseObject> CancelResponseAsync(string responseId, CancellationToken ct = default)

Types

Types are C# records defined in the LiterLlm namespace, serialized with System.Text.Json using snake_case naming policy.

ChatCompletionRequest

var request = new ChatCompletionRequest(
    Model: "gpt-4o-mini",
    Messages: [new UserMessage("Hello!")],
    MaxTokens: 256
);

ChatCompletionResponse

Property Type Description
Id string Response ID
Model string Model used
Choices Choice[] Completion choices
Usage Usage? Token usage

Error Handling

All errors derive from LlmException with numeric error codes:

Exception Code HTTP Status
InvalidRequestException 1400 400, 422
AuthenticationException 1401 401, 403
NotFoundException 1404 404
RateLimitException 1429 429
ProviderException 1500 5xx
StreamException 1600 --
SerializationException 1700 -- 
try
{
    var response = await client.ChatAsync(request);
}
catch (RateLimitException ex)
{
    Console.Error.WriteLine($"Rate limited: {ex.Message}");
}
catch (LlmException ex)
{
    Console.Error.WriteLine($"Error {ex.ErrorCode}: {ex.Message}");
}

Example

using LiterLlm;

await using var client = new LlmClient(
    apiKey: Environment.GetEnvironmentVariable("OPENAI_API_KEY")!);

var request = new ChatCompletionRequest(
    Model: "gpt-4o-mini",
    Messages: [new UserMessage("Hello!")],
    MaxTokens: 256);

var response = await client.ChatAsync(request);
Console.WriteLine(response.Choices[0].Message.Content);