PromptVit 1.1.1

PromptVit

Lightweight multi-provider LLM client for .NET

PromptVit is a modern, lightweight C# library that simplifies calls to major generative AI inference providers (OpenAI, Google Gemini, HuggingFace, Groq, etc.) from your .NET applications. It offers a unified, intuitive, and high-performance API to:

• Send simple or complex prompts
• Handle streaming responses (Server-Sent Events)
• Use tool/function calling in a standardized way
• Easily switch between different providers without changing your business logic

✨ Why PromptVit ?

• Ultra-lightweight: no unnecessary dependencies, no heavy SDKs
• Unified API: same code for OpenAI, Gemini, Claude, Groq…
• Native streaming: full support for real-time responses
• Standardized tool calls: JSON Schema format compatible with most providers
• Easy switching: switch providers in one line of configuration
• History persistence: save and reload chat history across sessions with a single call
• Strongly typed: well-typed messages, tools, responses, and errors
• High-performance: designed for production use (retry, timeout, logging)

Installation

dotnet add package PromptVit

Example

Simple chat

var aiClient = PromptVitFactory.CreateGoogleAIStudioClient("your_token", "gemini-3.1-flash-lite-preview");

aiClient.SetSystemPrompt("You are an AI assistant expert in cooking. You can answer to every questions related to cooking but cannot answer to anything else.");

var userInput = "How to make a Tiramisu ?";
var response = await aiClient.Invoke(userInput);

Chat with stream

var aiClient = PromptVitFactory.CreateGoogleAIStudioClient("your_token", "gemini-3.1-flash-lite-preview");
aiClient.OnStreamChunkReceived += async (s, chunk) =>
{
    Console.Write(chunk);
};

aiClient.SetSystemPrompt("You are an AI assistant expert in cooking. You can answer to every questions related to cooking but cannot answer to anything else.");

var userInput = "How to make a Tiramisu ?";
var response = await aiClient.Invoke(userInput, true);

Tool calling

var aiClient = PromptVitFactory.CreateGoogleAIStudioClient("your_token", "gemini-3.1-flash-lite-preview");

aiClient.SetSystemPrompt("You are an AI assistant that can give live weather information. You always use available tools to answer to user request.");

aiClient.SetTools([
    new AITool<string, WeatherArgs>("getWeather", "Retrieves weather for a given location",
    [
        new AIToolParameter("location", "a city location (example: Paris, France)", "string")
    ],
    async (arg) => await GetWeather(arg.Location))
]);

var userInput = "What is the weather in Madrid ?";
var response = await aiClient.Invoke(userInput);

static async Task<string> GetWeather(string location)
{
    return "the temperature is 15°C with a clear sky for the whole day";
}

class WeatherArgs
{
    public string Location { get; set; }
}

Save / load chat history

var aiClient = PromptVitFactory.CreateGoogleAIStudioClient("your_token", "gemini-3.1-flash-lite-preview");

aiClient.SetSystemPrompt("You are an AI assistant expert in cooking.");

await aiClient.Invoke("How to make a Tiramisu ?");
await aiClient.SaveHistoryAsync("history.txt");

// Later, reload the conversation context
await aiClient.LoadHistoryAsync("history.txt");
await aiClient.Invoke("What was the first dessert I asked about?");