README.md

ChatGPT.Net - .Net Library for ChatGPT [Discord]

Other versions [NodeJS Version]

Check the new Google Bard Chatbot!

The ChatGPT.Net is a C# library for ChatGPT using official OpenAI API that allows developers to access ChatGPT, a chat-based large language model. With this API, developers can send queries to ChatGPT and receive responses in real-time, making it easy to integrate ChatGPT into their own applications.

using ChatGPT.Net;

// ChatGPT Official API
var bot = new ChatGpt("<API_KEY>");

var response = await bot.Ask("What is the weather like today?");
Console.WriteLine(response);

Table of Contents

• Features
• Getting Started
  • Usage
    • ChatGPT Official API
    • ChatGPT Unofficial API
• Configuration options
  • ChatGPT Official API
  • ChatGPT Unofficial API
• Examples
  • ChatGPT Console App
  • Use a different model
  • Using ChatGPT Official API For Free
• License

Features

• Easy to use.
• Using official OpenAI API.
• Supports both free and pro accounts.
• Supports multiple accounts, and multiple conversations.
• Support response streaming, so you can get response while the model is still generating it.

Getting Started

To install ChatGPT.Net, run the following command in the Package Manager Console:

Install-Package ChatGPT.Net

Alternatively, you can install it using the .NET Core command-line interface:

dotnet add package ChatGPT.Net

Usage

ChatGPT Official API

Here is a sample code showing how to use ChatGPT.Net:

using ChatGPT.Net;

// ChatGPT Official API
var bot = new ChatGpt("<API_KEY>");

// get response
var response = await bot.Ask("What is the weather like today?");
Console.WriteLine(response);

// stream response
await bot.AskStream(response => {
    Console.WriteLine(response);
}, "What is the weather like today?");

// get response for a specific conversation
var response = await bot.Ask("What is the weather like today?", "conversation name");
Console.WriteLine(response);

// stream response for a specific conversation
await bot.AskStream(response => {
    Console.WriteLine(response);
}, "What is the weather like today?", "conversation name");

// Set a system message
bot.SetConversationSystemMessage("conversation name", "You are a helpful assistant that provides clear and concise answers to questions.");

Some models can understand both text and images. To use this feature, you need to pass a list of content items that includes both text and images.

// To stream responses with both image and text input, create a list of content items
var contentItems = new List<ChatGptMessageContentItem>();

// Each content item can be either "Text" type or "Image" type. Let's add text to inquire about the image
contentItems.Add(new ChatGptMessageContentItem()
{
    Type = ChatGptMessageContentType.TEXT,
    Text = "what is this image about?"
});

// Now, create another content item of "Image" type
var contentItemWithImage = new ChatGptMessageContentItem()
{
    Type = ChatGptMessageContentType.IMAGE
};
// Set image by path
contentItemWithImage.SetImageFromFile("<path-to-file>");
// Or set image by Url
contentItemWithImage.SetImageFromUrl("https://path-to-image.com/image.png");
// Or set base64 image url
contentItemWithImage.SetImageFromUrl("data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAA.....");

contentItems.Add(contentItemWithImage);

// Only specific models understand image input. Therefore, if you are using 'gpt-3.5-turbo' or similar model, switch or override the model before calling 'AskStream'. More details about `ChatGptOptions` (config) are covered in the next section.
config.Model = ChatGptModels.GPT_4_Vision_Preview;

var response = await bot.AskStream(response => {
    Console.WriteLine(response);
}, contentItems, "conversation name");

ChatGPT Unofficial API

Here is a sample code showing how to integrate (chat.openai.com) with your applications using ChatGPT.Net:

using ChatGPT.Net;

// ChatGPT Official API
var bot = new ChatGptUnofficial("<SESSION_TOKEN>");

// get response
var response = await bot.Ask("What is the weather like today?");
Console.WriteLine(response);

// stream response
await bot.AskStream(response => {
    Console.WriteLine(response);
}, "What is the weather like today?");

// get response for a specific conversation
var response = await bot.Ask("What is the weather like today?", "conversation name");
Console.WriteLine(response);

// stream response for a specific conversation
await bot.AskStream(response => {
    Console.WriteLine(response);
}, "What is the weather like today?", "conversation name");

Configuration options

ChatGPT Official API

ChatGptOptions
{
    string BaseUrl; // Default: https://api.openai.com
    string Model; // Default: gpt-3.5-turbo
    double Temperature; // Default: 0.9;
    double TopP; // Default: 1.0;
    long MaxTokens; // Default: 64;
    string[]? Stop; // Default: null;
    double PresencePenalty; // Default: 0.0;
    double FrequencyPenaltyl; // Default: 0.0;
}

ChatGPT Unofficial API

ChatGptUnofficialOptions
{
    string BaseUrl; // Default: https://api.pawan.krd
    string Model; // Default: text-davinci-002-render-sha
}

Examples

ChatGPT Console App

This is a simple console app that uses ChatGPT.Net to interact with ChatGPT.

using ChatGPT.Net;

// ChatGPT Official API
var bot = new ChatGpt("<API_KEY>");

var prompt = string.Empty;

while (true)
{
    Console.Write("You: ");
    prompt = Console.ReadLine();
    if (prompt is null) break;
    if (string.IsNullOrWhiteSpace(prompt)) break;
    if (prompt == "exit") break;
    Console.Write("ChatGPT: ");
    await bot.AskStream(Console.Write, prompt, "default");
    Console.WriteLine();
}

Use a different model

You can use a different model by passing the model name to the constructor.

var bot = new ChatGpt("<API_KEY>", new ChatGptOptions
{
    Model = "text-davinci-002-render-paid"
});

Using ChatGPT Official API For Free

you can use ChatGPT Official API by setting the base url to a free reverse proxy server like ChatGPT Free Reverse Proxy

var bot = new ChatGpt("<API_KEY>", new ChatGptOptions
{
    BaseUrl = "https://api.pawan.krd"
});

License

This project is licensed under the MIT License - see the LICENSE file for details