GenerativeAI.Gemini 1.0.2

GenerativeAI.Gemini

A comprehensive .NET 9 library for seamless integration with Google's Gemini generative AI models. This package provides strongly-typed clients, dependency injection support, structured response handling, and advanced features like tool integration and schema-based generation.

Features

• Strongly-typed API - Full IntelliSense support with comprehensive type safety
• Dependency Injection - Native ASP.NET Core DI integration
• Flexible Prompt Building - Support for text, files, and mixed content
• Structured Responses - Automatic JSON schema generation and deserialization
• Tool Integration - Built-in support for code execution, web search, and URL context
• Async/Await - Modern async programming patterns
• Configuration Binding - Seamless integration with .NET configuration system

Installation

Add the NuGet package to your project:

dotnet add package GenerativeAI.Gemini

Getting Started

Option 1: Using Dependency Injection (Recommended)

Register Services with API Key

You can register the Gemini client in your DI container using an API key and (optionally) a model:

services.AddGemini("your-api-key", "optional-model-name");

Or bind from configuration (example for appsettings.json):

{
  "Gemini": {
    "ApiKey": "your-api-key",
    "Model": "optional-model-name"
  }
}

2. Create and Execute Prompts

Create a prompt using IGeminiClient, then add text or inline data as needed:

public class Example
{
    private readonly IGeminiClient _geminiClient;

    public Example(IGeminiClient geminiClient)
    {
        _geminiClient = geminiClient;

        // Alternative: Create GeminiClient directly without DI
        // _geminiClient = new GeminiClient("your-api-key", GeminiModel.Gemini2_5Pro);
    }

    public async Task<IEnumerable<string>> GetResponseAsync()
    {
        // Create a prompt
        var prompt = _geminiClient.CreatePrompt();

        // Add text to the prompt
        prompt.AddText("You are a helpful assistant. Answer the question: What is the capital of France?");

        // Optionally, add inline data (e.g., an image as base64)
        // prompt.AddInlineData(base64ImageString, "image/png");

        // Execute the model and get a plain text response
        var response = await _geminiClient.GenerateTextAsync(prompt);

        // Extract all the text parts
        var textParts = response.TextParts.Select(part => part.text);

        return textParts;
    }
}

3. Advanced Prompt Options

You can customize prompt options such as temperature, top-k, top-p, and max output tokens by passing a GeminiPromptOptions instance to CreatePrompt:

// Example of customizing prompt options
var options = new GeminiPromptOptions
{
    Temperature = 0.7,
    TopK = 40,
    TopP = 0.9,
    MaxTokens = 150
};

var prompt = _geminiClient.CreatePrompt(options);

License

This project is licensed under the MIT License.

Important Notes

• This library is in preview and may change in future releases.
• Gemini API usage may incur costs.