AnthropicClient 0.1.0

AnthropicClient

This library for the Anthropic API is meant to simplify development in C# for Anthropic users.

[!NOTE] This is an unofficial SDK for the Anthropic API. It was not built in consultation with Anthropic or any member of their organization.

This SDK was developed independently using existing libraries and the Anthropic API documentation as the starting point with the intention of making development of integrations done in C# with Anthropic quicker and more convenient.

[!NOTE]
This client library is heavily inspired by the Anthropic.SDK library. I chose to create a new library because I wanted to handle streaming and tool calling differently as well as have control over the client library as I plan to use it to build a connector for SemanticKernel. However if you are looking for a client library the Anthropic.SDK is a great place to start.

📝 Issues

If you encounter any issues while using this library please open an issue here.

📜 License

This library is licensed under the MIT License and is free to use and modify.

📝 Contributing

If you would like to contribute to this library please open a pull request here.

🛠️ Dependencies

Microsoft.Bcl.AsyncInterfaces

Used to support async interfaces when streaming messages

System.Text.Json

Used for JSON serialization and deserialization

💾 Installation

Install the package from NuGet using the following command:

dotnet add package AnthropicClient

🔑 API Key

In order to use the Anthropic API you will need an API key. You can get one by signing up at Anthropic. Please keep your API key secure and do not share it with others. Be mindful of where you store your API key and do not commit it to a public repository.

👨🏻‍💻 Start Coding

AnthropicApiClient

The most common way to use the SDK is to create an AnthropicApiClient instance and call its methods. Its constructor requires two parameters:

• apiKey - your Anthropic API key
• httpClient - an HttpClient instance. You can configure and customize the HttpClient instance as needed. This library however will perform the necessary configuration to work with the Anthropic API. Such as setting the base address and adding the proper headers.

[!NOTE] This library does not manage the lifecycle of the HttpClient instance. You should create and manage the lifecycle of the HttpClient instance in your application.

It is best practice to read the API key from a secure location such as a configuration file or environment variable. For example using the appsettings.json file:

{
  "AnthropicApiKey": "YOUR_API"
}

Example constructing an AnthropicApiClient instance:

using AnthropicClient;
using Microsoft.Extensions.Configuration;

var configuration = new ConfigurationBuilder()
  .AddJsonFile("appsettings.json")
  .Build();

var apiKey = configuration["AnthropicApiKey"];

var client = new AnthropicApiClient(apiKey, new HttpClient());

IAnthropicApiClient

The library does expose an interface IAnthropicApiClient that can be used for dependency injection and testing. The interface is implemented by the AnthropicApiClient class.

Full API Documentation

This library was developed to make using the Anthropic API easier within a .NET application. If you are looking for the full API documentation you can find it at Anthropic API Documentation.

Usage

The primary use case for working with the Anthropic API is to create a message in response to a request that includes one or more other messages. The created message can then be received either as a complete response or a stream of events. This can be used to create a conversation between the caller and Anthropic's AI models and/or to use Anthropic's AI models to perform a task.

[!NOTE] The following examples assume that you have already created an instance of the AnthropicApiClient class named client. You can also find these snippets in the examples directory.

Create a message

The AnthropicApiClient exposes a single method named CreateMessageAsync that can be used to create a message. The method requires a MessageRequest or a StreamMessageRequest instance as a parameter. The MessageRequest class is used to create a message whose response is not streamed and the StreamMessageRequest class is used to create a message whose response is streamed. The MessageRequest instance's properties can be set to configure how the message is created.

Non-Streaming

using AnthropicClient;
using AnthropicClient.Models;

var response = await client.CreateMessageAsync(new MessageRequest(
  AnthropicModels.Claude3Haiku,
  [
    new(
      MessageRole.User, 
      [new TextContent("Please write a haiku about the ocean.")]
    )
  ]
));

if (response.IsSuccess is false)
{
  Console.WriteLine($"Failed to create message");
  Console.WriteLine($"Error Type: {0}", response.Error.Error.Type);
  Console.WriteLine($"Error Message: {0}", response.Error.Error.Message);
  return;
}

foreach (var content in response.Value.Content)
{
  switch (content)
  {
    case TextContent textContent:
      Console.WriteLine(textContent.Text);
      break;
  }
}

Streaming

Anthropic uses Server-Sent Events (SSE) to stream messages. The possible events and the format of those events are documented in the Anthropic API Documentation. This library provides a way to consume them after they have been deserialized into strongly-typed C# objects that are returned in an IAsyncEnumerable collection.

This allows you to consume the events as they are received and process them in the way that best fits your use case. The following example demonstrates how to consume the streamed events and build up the complete text response from the model.

using AnthropicClient;
using AnthropicClient.Models;

var events = client.CreateMessageAsync(new StreamMessageRequest(
  AnthropicModels.Claude3Haiku,
  [
    new(
      MessageRole.User, 
      [new TextContent("Please write a haiku about the ocean.")]
    )
  ]
));

var msgBuilder = new StringBuilder();

await foreach (var e in events)
{
  switch (e.Data)
  {
    case var data when data is ContentDeltaEventData contentData:
      switch (contentData.Delta)
      {
        case var delta when delta is TextDelta textDelta:
          msgBuilder.Append(textDelta.Text);
          break;
      }
      break;
  }
}

Console.WriteLine(msgBuilder.ToString());

Message Complete Event

This library also provides a custom message_complete event that is yielded when all the message's events have been received. This event is not part of Anthropic's SSE events but is provided to allow for easier consumption of the entire message response if desired and make it easier to implement built-in tool calling.

using AnthropicClient;
using AnthropicClient.Models;

var events = client.CreateMessageAsync(new StreamMessageRequest(
  AnthropicModels.Claude3Haiku,
  [
    new(
      MessageRole.User, 
      [new TextContent("Please write a haiku about the ocean.")]
    )
  ]
));

MessageResponse? response = null;

await foreach (var e in events)
{
  switch (e.Data)
  {
    case var data when data is MessageCompleteEventData msgData:
      response = msgData.Message;
      break;
  }
}

var textContent = response?.Content
  .OfType<TextContent>()
  .Aggregate(new StringBuilder(), (sb, c) => sb.Append(c.Text))
  .ToString();

Console.WriteLine(textContent);

Tool Use

Anthropic's models support the use of tools to perform tasks. This allows the models to interact with external client-side tools that can perform actions the models cannot do natively. This gives you the ability to further extend the model's abilities with your own custom tools. This feature is covered in depth in Anthropic's API Documentation. This library aims to make using tools convenient by allowing you to create, provide, and call tools from within your application by leveraging the reflection capabilities of C#.

[!NOTE] All tools are user provided. The models do no not have access to any built-in server-side tools.

Create a tool

You can create a tool in 4 different ways and then provide that tool when creating a message.

1. Create a tool from a class
2. Create a tool from a static method
3. Create a tool from an instance method
4. Create a tool from a delegate

Create a tool from a class

When creating a tool from a class the class must implement the ITool interface.

using AnthropicClient.Models;

class GetWeatherTool : ITool
{
  public string Name => "Get Weather";

  public string Description => "Get the weather for a location in the specified units";

  public MethodInfo Function => typeof(GetWeatherTool).GetMethod(nameof(GetWeather))!;

  public static string GetWeather(string location, string units)
  {
    return $"The weather in {location} is 72 degrees {units}";
  }
}

var getWeatherTool = Tool.CreateFromClass<GetWeatherTool>();

Create a tool from a static method

When creating a tool from a static method the method must be public and static.

using AnthropicClient.Models;

class GetWeatherTool
{
  public static string GetWeather(string location)
  {
    return $"The weather in {location} is 72 degrees Fahrenheit";
  }
}

var getWeatherTool = Tool.CreateFromStaticMethod(
  "Get Weather", 
  "Get the weather for a location in the specified units", 
  typeof(GetWeatherTool), 
  nameof(GetWeatherTool.GetWeather)
);

Create a tool from an instance method

When creating a tool from an instance method the method must be public and non-static.

using AnthropicClient.Models;

class GetWeatherTool
{
  public string GetWeather(string location)
  {
    return $"The weather in {location} is 72 degrees Fahrenheit";
  }
}

var toolInstance = new GetWeatherTool();

var getWeatherTool = Tool.CreateFromInstanceMethod(
  "Get Weather", 
  "Get the weather for a location in the specified units", 
  toolInstance,
  nameof(toolInstance.GetWeather)
);

Create a tool from a delegate

When creating a tool from a delegate the delegate must be a Func<TResult>, Func<T, TResult>, or Func<T1, T2, TResult>. If you need to create a tool from a delegate that takes more than 2 parameters you should create a complex type and pass that as the parameter.

using AnthropicClient.Models;

var tool = (string location, string units) => $"The weather in {location} is 72 degrees {units}";

var getWeatherTool = Tool.CreateFromFunction(
  "Get Weather", 
  "Get the weather for a location in the specified units", 
  tool
);

Function Parameter Attribute

When you create a tool from one of the methods above and send it to Anthropic in your request a JSON representation of the tool is provided in the message. This JSON representation includes the name, description, and input schema of the tool. This information is used by Anthropic's models to discern if and when it should use a tool.

This library provides a FunctionParameterAttribute that can be used to provide additional information about the parameters of the tool. This information is used to provide a more detailed input schema for the tool.

using AnthropicClient.Models;

var tool = (
  [FunctionParameter(description: "The location of the weather being got", name: "Location", required: true)] 
  string location, 
  string units
) => $"The weather in {location} is 72 degrees {units}";

var getWeatherTool = Tool.CreateFromFunction(
  "Get Weather", 
  "Get the weather for a location in the specified units", 
  tool
);

Function Property Attribute

This library also provides a FunctionPropertyAttribute that can be used to provide additional information about the members of complex types used as parameters in the tool. This information is used to provide a more detailed input schema for the tool.

using AnthropicClient.Models;

class GetWeatherInput
{
  [FunctionProperty(
    description: "The location of the weather being got",
    required: true
  )]
  public string Location { get; } = string.Empty;

  [FunctionProperty(
    description: "The units to get the weather in",
    required: false,
    defaultValue: "Fahrenheit",
    possibleValues: ["Fahrenheit", "Celsius"]
  )]
  public string Units { get; } = "Fahrenheit";
}

  var tool = (GetWeatherInput input) => $"The weather in {input.Location} is 72 degrees {input.Units}";

  var getWeatherTool = Tool.CreateFromFunction(
    "Get Weather", 
    "Get the weather for a location in the specified units", 
    tool
  );

Call a tool

It is important to remember that while Anthropic's models do support tool use they don't actually have access to any built-in server-side tools. All tools are user provided. This means that while Anthropic's models can respond to a request to create a message with a request to use a tool that is all it is - a request. It is still up to the client to handle the tool request by calling the tool with the input provided by the model and then providing the result of that call back to the model.

This library aims to make this process convenient by allowing you to simply provide the tools you want Anthropic's models to consider for use when creating a message, receive the response, check if the response contains a tool call, and if it does invoke the tool to get the result.

[!NOTE] Anthropic's API expects requests to contain messages that alternate between the user and the assistant. In addition if you receive a tool use from the model the API expects you to respond with a message that contains the result of the tool call. The tool use content will always be from the assistant while the tool result will always be from the user.

using AnthropicClient;
using AnthropicClient.Models;

class GetWeatherTool : ITool
{
  public string Name => "Get Weather";

  public string Description => "Get the weather for a location in the specified units";

  public MethodInfo Function => typeof(GetWeatherTool).GetMethod(nameof(GetWeather))!;

  public static string GetWeather(string location, string units)
  {
    return $"The weather in {location} is 72 degrees {units}";
  }
}

List<Message> messages = [
  new(
    MessageRole.User, 
    [new TextContent("What is the weather in New York?")]
  )
];

List<Tool> tools = [Tool.CreateFromClass<GetWeatherTool>()];

var response = await client.CreateMessageAsync(new MessageRequest(
  AnthropicModels.Claude3Haiku,
  messages,
  tools: tools
));

if (response.IsSuccess is false)
{
  Console.WriteLine($"Failed to create message");
  Console.WriteLine($"Error Type: {0}", response.Error.Error.Type);
  Console.WriteLine($"Error Message: {0}", response.Error.Error.Message);
  return;
}

messages.Add(new(MessageRole.Assistant, response.Content));


foreach (var content in response.Value.Content)
{
  switch (content)
  {
    case TextContent textContent:
      Console.WriteLine(textContent.Text);
      break;
    case ToolUseContent toolUseContent:
      Console.WriteLine(toolUseContent.Name);
      break;
  }
}

if (response.Value.ToolCall is not null)
{
  var toolCallResult = await response.Value.ToolCall.InvokeAsync<string>();
  string toolResultContent;

  if (toolCallResult.IsSuccess && toolCallResult.Value is not null)
  {
    Console.WriteLine(toolCallResult.Value);
    toolResultContent = toolCallResult.Value;
  }
  else
  {
    Console.WriteLine(toolCallResult.Error.Message);
    toolResultContent = toolCallResult.Error.Message;
  }

  messages.Add(
    new(
      MessageRole.User, 
      [
        new ToolResultContent(
          response.Value.ToolCall.ToolUse.Id, 
          toolResultContent
        )
      ]
    )
  );
}

var finalResponse = await client.CreateMessageAsync(new MessageRequest(
  AnthropicModels.Claude3Haiku,
  messages,
  tools: tools
));

if (finalResponse.IsSuccess is false)
{
  Console.WriteLine($"Failed to create message");
  Console.WriteLine($"Error Type: {0}", finalResponse.Error.Error.Type);
  Console.WriteLine($"Error Message: {0}", finalResponse.Error.Error.Message);
  return;
}

foreach (var content in finalResponse.Value.Content)
{
  switch (content)
  {
    case TextContent textContent:
      Console.WriteLine(textContent.Text);
      break;
  }
}

If an exception is thrown while invoking the tool the InvokeAsync method will return a ToolCallResult with the exception contained in the Error property.

[!NOTE] The InvokeAsync method does accept a generic type parameter that can be used to specify the type of the Value property of the ToolCallResult. If it is not specified it will be an object.

Call a tool in streamed message

Tool calling is also supported when streaming the message response. The following example demonstrates how you can handle a tool call in a streamed message response.

using AnthropicClient;
using AnthropicClient.Models;

var tool = (string location, string units) => $"The weather in {location} is 72 degrees {units}";

var messages = [
  new(
    MessageRole.User, 
    [new TextContent("What is the weather in New York?")]
  )
];

var tools = [Tool.CreateFromFunction(
  "Get Weather", 
  "Get the weather for a location in the specified units", 
  tool
)];

var events = client.CreateMessageAsync(new StreamMessageRequest(
  AnthropicModels.Claude3Haiku,
  messages,
  tools: tools
));

MessageResponse? response = null;

await foreach (var e in events)
{
  switch (e.Data)
  {
    case var data when data is MessageCompleteEventData msgData:
      response = msgData.Message;
      break;
  }
}

if (response is null)
{
  Console.WriteLine("Failed to get message response");
  return;
}

messages.Add(new(MessageRole.Assistant, response.Content));

if (response?.ToolCall is not null)
{
  var toolCallResult = await response.ToolCall.InvokeAsync<string>();
  string toolResultContent;

  if (toolCallResult.IsSuccess && toolCallResult.Value is not null)
  {
    toolResultContent = toolCallResult.Value;
  }
  else
  {
    toolResultContent = toolCallResult.Error.Message;
  }

  messages.Add(
    new(
      MessageRole.User, 
      [
        new ToolResultContent(
          response.ToolCall.ToolUse.Id, 
          toolResultContent
        )
      ]
    )
  );
}

var finalResponse = await client.CreateMessageAsync(new MessageRequest(
  AnthropicModels.Claude3Haiku,
  messages,
  tools: tools
));

if (finalResponse.IsSuccess is false)
{
  Console.WriteLine($"Failed to create message");
  Console.WriteLine($"Error Type: {0}", finalResponse.Error.Error.Type);
  Console.WriteLine($"Error Message: {0}", finalResponse.Error.Error.Message);
  return;
}

foreach (var content in finalResponse.Value.Content)
{
  switch (content)
  {
    case TextContent textContent:
      Console.WriteLine(textContent.Text);
      break;
  }
}

If you do find that you need more control over how exactly provided tools are called and how the result of those tools are returned you can avoid using the InvokeAsync method and instead use the Tool and ToolUse properties of the ToolCall instance to implement your own solution.

System Prompt

Anthropic's models support the use of system prompts to provide additional context to the user. This can be used to provide additional information to the user or to ask for additional information from the user. This feature is covered in depth in Anthropic's API Documentation. This library aims to make using system prompts convenient by allowing you to provide the system prompts you want Anthropic's models to consider for use when creating a message.

System Message

You can create a system prompt by providing a string as the system parameter in the MessageRequest or StreamMessageRequest constructor.

using AnthropicClient;
using AnthropicClient.Models;

var response = await client.CreateMessageAsync(new MessageRequest(
  AnthropicModels.Claude3Haiku,
  [
    new(
      MessageRole.User, 
      [new TextContent("Please write a haiku about the ocean.")]
    )
  ],
  system: "You are a internationally renowned poet. You excel at writing haikus.
));

if (response.IsSuccess is false)
{
  Console.WriteLine($"Failed to create message");
  Console.WriteLine($"Error Type: {0}", response.Error.Error.Type);
  Console.WriteLine($"Error Message: {0}", response.Error.Error.Message);
  return;
}

foreach (var content in response.Value.Content)
{
  switch (content)
  {
    case TextContent textContent:
      Console.WriteLine(textContent.Text);
      break;
  }
}

System Messages

You can create a more complex system prompt by providing a List<TextContent> as the systemMessages parameter in the MessageRequest or StreamMessageRequest constructor.

using AnthropicClient;
using AnthropicClient.Models;

var response = await client.CreateMessageAsync(new MessageRequest(
  AnthropicModels.Claude3Haiku,
  [
    new(
      MessageRole.User, 
      [new TextContent("Please write a haiku about the ocean.")]
    )
  ],
  systemMessages: [
    new TextContent("You are a internationally renowned poet. You excel at writing haikus."),
    new TextContent("You have been asked to write a haiku about the ocean.")
  ]
));

if (response.IsSuccess is false)
{
  Console.WriteLine($"Failed to create message");
  Console.WriteLine($"Error Type: {0}", response.Error.Error.Type);
  Console.WriteLine($"Error Message: {0}", response.Error.Error.Message);
  return;
}

foreach (var content in response.Value.Content)
{
  switch (content)
  {
    case TextContent textContent:
      Console.WriteLine(textContent.Text);
      break;
  }
}

Prompt Caching

Anthropic has recently introduced a feature called Prompt Caching that allows you to cache all or part of the prompt you send to the model. This can be used to improve the performance of your application by reducing latency and token usage. This feature is covered in depth in Anthropic's API Documentation.

[!NOTE] This feature is in beta and requires you to set an anthropic-beta header on your requests to use it. The value of the header should be prompt-caching-2024-07-31.

When using this library you can opt-in to prompt caching by adding the required header to the HttpClient instance you provide to the AnthropicApiClient constructor.

using AnthropicClient;

var httpClient = new HttpClient();
httpClient.DefaultRequestHeaders.Add("anthropic-beta", "prompt-caching-2024-07-31");

var client = new AnthropicApiClient(apiKey, httpClient);

Prompt caching can be used to cache all parts of the prompt including system messages, user messages, and tools. You should refer to the Anthropic API Documentation for specifics on limitations and requirements for using prompt caching. This library aims to make using prompt caching convenient and give you complete control over what parts of the prompt are cached. Currently there is only one type of cache control available - EphemeralCacheControl.

Caching System Messages

System messages can be cached by providing a List<TextContent> as the systemMessages parameter in the MessageRequest or StreamMessageRequest constructor and having one or more of the TextContent instances have the CacheControl property set.

using AnthropicClient;
using AnthropicClient.Models;

var response = await client.CreateMessageAsync(new MessageRequest(
  AnthropicModels.Claude3Haiku,
  [
    new(
      MessageRole.User, 
      [new TextContent("Please write a haiku about the ocean.")]
    )
  ],
  systemMessages: [
    new TextContent("You are a internationally renowned poet. You excel at writing haikus. Please use the following as examples."),
    new TextContent(exampleHaikus, new EphemeralCacheControl())
  ]
));

if (response.IsSuccess is false)
{
  Console.WriteLine($"Failed to create message");
  Console.WriteLine($"Error Type: {0}", response.Error.Error.Type);
  Console.WriteLine($"Error Message: {0}", response.Error.Error.Message);
  return;
}

foreach (var content in response.Value.Content)
{
  switch (content)
  {
    case TextContent textContent:
      Console.WriteLine(textContent.Text);
      break;
  }
}

Caching User Messages

User messages can be cached by providing a List<Content> as the messages parameter in the MessageRequest or StreamMessageRequest constructor and having one or more of the Content instances have the CacheControl property set.

using AnthropicClient;
using AnthropicClient.Models;

var response = await client.CreateMessageAsync(new MessageRequest(
  AnthropicModels.Claude3Haiku,
  [
    new(
      MessageRole.User, 
      [
        new TextContent("Please write a haiku about the ocean. Here are some examples of haikus I like."),
        new TextContent(exampleHaikus, new EphemeralCacheControl())
      ]
    ),
  ]
));

if (response.IsSuccess is false)
{
  Console.WriteLine($"Failed to create message");
  Console.WriteLine($"Error Type: {0}", response.Error.Error.Type);
  Console.WriteLine($"Error Message: {0}", response.Error.Error.Message);
  return;
}

foreach (var content in response.Value.Content)
{
  switch (content)
  {
    case TextContent textContent:
      Console.WriteLine(textContent.Text);
      break;
  }
}

Caching Tools

Tools can be cached by providing a List<Tool> as the tools parameter in the MessageRequest or StreamMessageRequest constructor and having one or more of the Tool instances have the CacheControl property set. This property can be set after the tool is created manually or by using one of the static methods on the Tool class.

using AnthropicClient;
using AnthropicClient.Models;

var tool = (string location, string units) => $"The weather in {location} is 72 degrees {units}";

var getWeatherTool = Tool.CreateFromFunction(
  "Get Weather", 
  "Get the weather for a location in the specified units", 
  tool,
  new EphemeralCacheControl()
);

var response = await client.CreateMessageAsync(new MessageRequest(
  AnthropicModels.Claude3Haiku,
  [
    new(
      MessageRole.User, 
      [new TextContent("What is the weather in New York?")]
    )
  ],
  tools: [
    // Lots of other tools
    // ...
    getWeatherTool
  ]
));

if (response.IsSuccess is false)
{
  Console.WriteLine($"Failed to create message");
  Console.WriteLine($"Error Type: {0}", response.Error.Error.Type);
  Console.WriteLine($"Error Message: {0}", response.Error.Error.Message);
  return;
}

foreach (var content in response.Value.Content)
{
  switch (content)
  {
    case TextContent textContent:
      Console.WriteLine(textContent.Text);
      break;
    case ToolUseContent toolUseContent:
      Console.WriteLine(toolUseContent.Name);
      break;
  }
}