LlmTornado

Getting Started ​

Welcome to the LlmTornado documentation! This comprehensive guide will walk you through installation, setup, and your first AI-powered application using LlmTornado.

Why LlmTornado? ​

LlmTornado is a powerful, flexible C# library that simplifies AI integration by:

  • Supporting 15+ AI providers with a unified API
  • Offering agentic frameworks for complex workflows
  • Providing multimodal capabilities (text, images, audio, video)
  • Enabling MCP integration for standardized tool communication
  • Ensuring enterprise readiness with comprehensive error handling

Installation ​

Core Package ​

Install the main LlmTornado package via NuGet:

bash
dotnet add package LlmTornado

Or using Package Manager Console:

powershell
Install-Package LlmTornado

Optional Addons ​

Enhance LlmTornado with additional packages:

bash
# Agentic framework for building autonomous AI agents
dotnet add package LlmTornado.Agents

# Model Context Protocol (MCP) integration
dotnet add package LlmTornado.Mcp

# Productivity and quality of life enhancements
dotnet add package LlmTornado.Contrib

# Vector database support
dotnet add package LlmTornado.VectorDatabases
dotnet add package LlmTornado.VectorDatabases.ChromaDB

Quick Start ​

Basic Chat Completion ​

Get started with a simple chat completion using OpenAI:

csharp
using LlmTornado;
using LlmTornado.Chat;
using LlmTornado.Chat.Models;

// Initialize with your API key
TornadoApi api = new TornadoApi(new[]
{
    new ProviderAuthentication(LLmProviders.OpenAi, "YOUR_OPENAI_API_KEY")
});

// Create and execute a chat request
ChatResult? result = await api.Chat.CreateChatCompletion(new ChatRequest
{
    Model = ChatModel.OpenAi.Gpt4.Turbo,
    Messages = [
        new ChatMessage(ChatMessageRoles.System, "You are a helpful assistant."),
        new ChatMessage(ChatMessageRoles.User, "What is 2+2?")
    ]
});

// Display the response
Console.WriteLine(result?.Choices?[0].Message?.Content);
// Output: "2+2 equals 4."

Using Conversations ​

For more natural interactions, use the conversation API:

csharp
using LlmTornado;
using LlmTornado.Chat;
using LlmTornado.Chat.Models;

TornadoApi api = new TornadoApi("YOUR_OPENAI_API_KEY");

// Create a conversation
Conversation conversation = api.Chat.CreateConversation(new ChatRequest
{
    Model = ChatModel.OpenAi.Gpt35.Turbo
});

conversation.AddSystemMessage("You are a helpful assistant.");

// Add messages and get responses
conversation.AddUserMessage("What is async/await in C#?");
string? response = await conversation.GetResponse();
Console.WriteLine(response);

// Continue the conversation
conversation.AddUserMessage("Can you show me an example?");
response = await conversation.GetResponse();
Console.WriteLine(response);

Multi-Provider Support ​

Switch between providers effortlessly:

csharp
// Initialize with multiple providers
TornadoApi api = new TornadoApi([
    new (LLmProviders.OpenAi, "OPENAI_KEY"),
    new (LLmProviders.Anthropic, "ANTHROPIC_KEY"),
    new (LLmProviders.Google, "GOOGLE_KEY"),
    new (LLmProviders.Groq, "GROQ_KEY")
]);

// Use OpenAI
Conversation openAiChat = api.Chat.CreateConversation(ChatModel.OpenAi.Gpt4.O);
await openAiChat.AppendUserInput("Hello!").GetResponse();

// Switch to Anthropic
Conversation claudeChat = api.Chat.CreateConversation(ChatModel.Anthropic.Claude37.Sonnet);
await claudeChat.AppendUserInput("Hello!").GetResponse();

// Use Google Gemini
Conversation geminiChat = api.Chat.CreateConversation(ChatModel.Google.Gemini.Gemini2Flash001);
await geminiChat.AppendUserInput("Hello!").GetResponse();

Configuration ​

API Keys ​

Option 1: Direct Initialization ​

csharp
TornadoApi api = new TornadoApi("YOUR_API_KEY");

Option 2: Multiple Providers ​

csharp
TornadoApi api = new TornadoApi([
    new (LLmProviders.OpenAi, "OPENAI_KEY"),
    new (LLmProviders.Anthropic, "ANTHROPIC_KEY"),
    new (LLmProviders.Google, "GOOGLE_KEY")
]);

Option 3: Environment Variables ​

csharp
// Set environment variable
Environment.SetEnvironmentVariable("OPENAI_API_KEY", "your-key");

// Use without explicit key
TornadoApi api = new TornadoApi(LLmProviders.OpenAi);

Custom Endpoints ​

For self-hosted or custom providers:

csharp
// Ollama example
TornadoApi api = new TornadoApi(new Uri("http://localhost:11434"));

Conversation chat = api.Chat.CreateConversation(new ChatModel("llama3:8b"));
await chat.AppendUserInput("Hello!").GetResponse();

Common Use Cases ​

1. Simple Q&A Bot ​

csharp
TornadoApi api = new TornadoApi("YOUR_API_KEY");

while (true)
{
    Console.Write("You: ");
    string? userInput = Console.ReadLine();
    
    if (string.IsNullOrEmpty(userInput) || userInput == "exit")
        break;
    
    Conversation chat = api.Chat.CreateConversation(ChatModel.OpenAi.Gpt35Turbo);
    string? response = await chat.AppendUserInput(userInput).GetResponse();
    
    Console.WriteLine($"AI: {response}");
}

2. Streaming Responses ​

csharp
TornadoApi api = new TornadoApi("YOUR_API_KEY");

Conversation chat = api.Chat.CreateConversation(ChatModel.OpenAi.Gpt4.O);

await chat
    .AppendSystemMessage("You are a creative storyteller.")
    .AppendUserInput("Tell me a short story about a robot.")
    .StreamResponse(Console.Write);

3. Function Calling ​

csharp
// Define a tool function
string GetCurrentWeather(string location)
{
    return $"Weather in {location}: Sunny, 72°F";
}

TornadoApi api = new TornadoApi("YOUR_API_KEY");

Conversation chat = api.Chat.CreateConversation(new ChatRequest
{
    Model = ChatModel.OpenAi.Gpt4.O,
    Tools = [
        new Tool((string location, ToolArguments args) => GetCurrentWeather(location),
            "get_weather",
            "Get the current weather for a location")
    ]
});

await chat.AppendUserInput("What's the weather like in San Francisco?").GetResponseRich();

4. Structured Output ​

csharp
public class Person
{
    public string Name { get; set; }
    public int Age { get; set; }
    public string City { get; set; }
}

TornadoApi api = new TornadoApi("YOUR_API_KEY");

Conversation chat = api.Chat.CreateConversation(new ChatRequest
{
    Model = ChatModel.OpenAi.Gpt4.O,
    ResponseFormat = ChatRequestResponseFormats.Json
});

chat.AppendSystemMessage("Extract person info from the message. Return JSON with name, age, and city.");
chat.AppendUserInput("John is 30 years old and lives in New York.");

ChatRichResponse response = await chat.GetResponseRich();
Person? person = JsonConvert.DeserializeObject<Person>(response.Content);

5. Building an Agent ​

csharp
using LlmTornado.Agents;

TornadoApi api = new TornadoApi("YOUR_API_KEY");

TornadoAgent agent = new TornadoAgent(
    client: api,
    model: ChatModel.OpenAi.Gpt41.V41Mini,
    instructions: "You are a helpful coding assistant specialized in C#."
);

Conversation result = await agent.RunAsync("Explain the SOLID principles");
Console.WriteLine(result.Messages.Last().Content);

Next Steps ​

Now that you've got the basics, explore more advanced features:

Core Features ​

Advanced Topics ​

Troubleshooting ​

Common Issues ​

API Key Errors ​

Error: Invalid API key

Solution: Verify your API key is correct and has proper permissions

Model Not Found ​

Error: Model 'xyz' not found

Solution: Check the Feature Matrix for supported models

Rate Limiting ​

Error: Rate limit exceeded

Solution: Implement retry logic with exponential backoff or upgrade your API plan

Getting Help ​

  • Documentation: Browse the complete documentation in the sidebar
  • GitHub Issues: Report bugs or request features at GitHub
  • Examples: Check the Demo project for more examples

Additional Resources ​