Ollama SDK for .NET 🦙

Generated C# SDK based on Ollama OpenAPI specification and official docs using OpenApiGenerator.

Features 🔥

• Intuitive API client: Set up and interact with Ollama in just a few lines of code.
• API endpoint coverage: Support for all Ollama API endpoints including chats, embeddings, listing models, pulling and creating new models, and more.
• Real-time streaming: Stream responses directly to your application.
• Progress reporting: Get real-time progress feedback on tasks like model pulling.

Usage

Initializing

using var ollama = new OllamaApiClient();

var models = await ollama.Models.ListModelsAsync();

// Pulling a model and reporting progress
await foreach (var response in ollama.PullModelAsync("all-minilm", stream: true))
{
    Console.WriteLine($"{response.Status}. Progress: {response.Completed}/{response.Total}");
}
// or just pull the model and wait for it to finish
await ollama.Models.PullModelAndEnsureSuccessAsync("all-minilm");

// Generating an embedding
var embedding = await ollama.Embeddings.GenerateEmbeddingAsync(
    model: "all-minilm",
    prompt: "hello");

// Streaming a completion directly into the console
// keep reusing the context to keep the chat topic going
IList<long>? context = null;
var enumerable = ollama.Completions.GenerateCompletionAsync("llama3", "answer 5 random words", stream: true);
await foreach (var response in enumerable)
{
    Console.WriteLine($"> {response.Response}");
    
    context = response.Context;
}

var lastResponse = await ollama.Completions.GenerateCompletionAsync("llama3", "answer 123", stream: false, context: context).WaitAsync();
Console.WriteLine(lastResponse.Response);

var chat = ollama.Chat("mistral");
while (true)
{
    var message = await chat.SendAsync("answer 123");
    
    Console.WriteLine(message.Content);
    
    var newMessage = Console.ReadLine();
    await chat.Send(newMessage);
}

Credits

Icon and name were reused from the amazing Ollama project. The project was forked from this repository, after which automatic code generation was applied based on this OpenAPI specification (in the future it will be replaced by the official one, if one appears)