CompactifAI.Client 10.0.0

CompactifAI.Client

A .NET client library for the CompactifAI API. Easily integrate AI chat completions, text completions, and audio transcription into your .NET applications.

.NET Version Support

See VERSIONING.md for the complete versioning strategy.

Installation

# For .NET 10 (latest)
dotnet add package CompactifAI.Client --version 10.*

# For .NET 9
dotnet add package CompactifAI.Client --version 9.*

# For .NET 8
dotnet add package CompactifAI.Client --version 8.*

Quick Start

Simple Usage (without DI)

using CompactifAI.Client;

// Create client with API key
var client = new CompactifAIClient("your-api-key");

// Simple chat
var response = await client.ChatAsync("What is the capital of France?");
Console.WriteLine(response);
// Output: The capital of France is Paris.

// Chat with system prompt
var response = await client.ChatAsync(
    message: "Explain quantum computing",
    systemPrompt: "You are a helpful teacher. Explain concepts simply.",
    model: CompactifAIModels.DeepSeekR1_Slim);

Dependency Injection (ASP.NET Core)

appsettings.json:

{
  "CompactifAI": {
    "ApiKey": "your-api-key",
    "DefaultModel": "cai-llama-3-1-8b-slim"
  }
}

Program.cs:

using CompactifAI.Client.Extensions;

var builder = WebApplication.CreateBuilder(args);

// Option 1: From configuration
builder.Services.AddCompactifAI(builder.Configuration);

// Option 2: With API key directly
builder.Services.AddCompactifAI("your-api-key");

// Option 3: With full configuration
builder.Services.AddCompactifAI(options =>
{
    options.ApiKey = "your-api-key";
    options.DefaultModel = CompactifAIModels.Llama33_70B_Slim;
    options.TimeoutSeconds = 180;
});

Using in a service:

public class MyService
{
    private readonly ICompactifAIClient _client;

    public MyService(ICompactifAIClient client)
    {
        _client = client;
    }

    public async Task<string> GetAnswerAsync(string question)
    {
        return await _client.ChatAsync(question);
    }
}

Features

Chat Completions

using CompactifAI.Client;
using CompactifAI.Client.Models;

// Full control with ChatCompletionRequest
var request = new ChatCompletionRequest
{
    Model = CompactifAIModels.Llama4Scout_Slim,
    Messages = new List<ChatMessage>
    {
        ChatMessage.System("You are a helpful coding assistant."),
        ChatMessage.User("Write a function to calculate fibonacci numbers in C#")
    },
    Temperature = 0.7,
    MaxTokens = 1000
};

var response = await client.CreateChatCompletionAsync(request);
Console.WriteLine(response.Choices[0].Message?.Content);
Console.WriteLine($"Tokens used: {response.Usage?.TotalTokens}");

Text Completions

// Simple completion
var text = await client.CompleteAsync(
    prompt: "The quick brown fox",
    maxTokens: 50);

// Full control
var request = new CompletionRequest
{
    Model = CompactifAIModels.Llama31_8B_Slim,
    Prompt = "Once upon a time",
    Temperature = 0.9,
    MaxTokens = 200
};

var response = await client.CreateCompletionAsync(request);
Console.WriteLine(response.Choices[0].Text);

Audio Transcription

// Transcribe a file (MP3 recommended)
var text = await client.TranscribeFileAsync("audio.mp3", language: "en");
Console.WriteLine(text);

// Full control with TranscriptionRequest
var request = new TranscriptionRequest
{
    FileContent = await File.ReadAllBytesAsync("meeting.mp3"),
    FileName = "meeting.mp3",
    Model = CompactifAIModels.WhisperLargeV3,
    Language = "en"
};

var response = await client.TranscribeAsync(request);
Console.WriteLine($"Duration: {response.Duration}s");
Console.WriteLine($"Text: {response.Text}");

// Access segments for timestamps
foreach (var segment in response.Segments ?? new())
{
    Console.WriteLine($"[{segment.Start:F2}s - {segment.End:F2}s] {segment.Text}");
}

List Available Models

var models = await client.ListModelsAsync();
foreach (var model in models.Data)
{
    Console.WriteLine($"{model.Id} - {model.OwnedBy}");
}

// Get specific model info
var modelInfo = await client.GetModelAsync(CompactifAIModels.DeepSeekR1_Slim);
Console.WriteLine($"Parameters: {modelInfo.ParametersNumber}");

Available Models

Use the CompactifAIModels static class for strongly-typed model constants.

Error Handling

try
{
    var response = await client.ChatAsync("Hello!");
}
catch (CompactifAIException ex)
{
    Console.WriteLine($"API Error: {ex.StatusCode}");
    Console.WriteLine($"Message: {ex.Message}");
    Console.WriteLine($"Response: {ex.ResponseBody}");
}

Configuration Options

Getting an API Key

1. Subscribe through AWS Marketplace
2. Wait for approval (typically 24 hours)
3. Access your API key at MultiverseIAM Dashboard

License

MIT License