InstructSharp 1.0.0

InstructSharp

InstructSharp is your go-to .NET library for seamless, provider-agnostic LLM integrations. Whether you’re working with ChatGPT, Claude, Gemini, Grok, DeepSeek or LLaMA, InstructSharp wraps every call in a consistent QueryAsync<T> pattern and returns crisp, structured JSON objects 📦✨. Just define your POCO schema, pick your model, and you’re ready to 🚀:

• 🔌 Multi-provider support: OpenAI, Anthropic, Google, X.AI, DeepSeek & more
• 🔖 Type-safe outputs: Deserialize straight into your custom classes
• ⚙️ Fully configurable: Tweak temperature, max_tokens, and other settings
• 🛠️ Minimal setup: Install via NuGet & start querying in minutes

Crafted for clarity and speed, InstructSharp turns advanced AI workflows into everyday dev tasks 🎯👍.

Installation

dotnet add package InstructSharp

Or via Package Manager:

Install-Package InstructSharp

Supported Providers

• OpenAI ChatGPT (ChatGPTClient)
• Anthropic Claude (ClaudeClient)
• Google Gemini (GeminiClient)
• X.AI Grok (GrokClient)
• DeepSeek (DeepSeekClient)
• Meta LLaMA (via DeepInfra) (LLamaClient)

Each client implements the same QueryAsync<T> pattern and returns a LLMResponse<T>.

Quick Start

1. Define your output schema

   class QuestionAnswer
   {
       public string Question { get; set; }
       public string Answer   { get; set; }
   }
   

2. Instantiate a client

   // ChatGPT example
   var chat = new ChatGPTClient("YOUR_OPENAI_API_KEY");
   

3. Build a request

   var req = new ChatGPTRequest
   {
       Model       = ChatGPTModels.GPT41,
       Instruction = "Talk like a pirate.",
       Input       = "Are semicolons optional in JavaScript?"
   };
   

4. Send and receive structured output

   var res = await chat.QueryAsync<QuestionAnswer>(req);
   
   Console.WriteLine($"Q: {res.Result.Question}");
   Console.WriteLine($"A: {res.Result.Answer}");
   

5. Plain‐text fallback

   // For unstructured responses:
   var textRes = await chat.QueryAsync<string>(req);
   Console.WriteLine(textRes.Result);
   

Full Demo

using InstructSharp.Clients.ChatGPT;
using InstructSharp.Core;
using System.Text.Json;

// ...

var client = new ChatGPTClient("sk-...");
var request = new ChatGPTRequest {
    Model       = ChatGPTModels.GPT41,
    Instruction = "Talk like a pirate.",
    Input       = "What is 2 + 2?"
};

var response = await client.QueryAsync<QuestionAnswer>(request);

Console.WriteLine(JsonSerializer.Serialize(response.Result));
// → { "Question":"What is 2 + 2?","Answer":"Arr, 'tis four, matey!" }

Repeat the same steps for any other provider:

var claude = new ClaudeClient("YOUR_CLAUDE_KEY");
var gemini = new GeminiClient("YOUR_GOOGLE_KEY", GeminiModels.Flash25);
// …and so on…

Advanced

• Custom types: any POCO can be used as T.
• Temperature, max_tokens, etc. available via each provider’s request class.
• Error handling: exceptions thrown on non-2xx responses. Inspect response.Usage and response.AdditionalData for extra info.

Contributing

1. Fork the repo
2. Implement your feature / fix
3. Send a PR