Getting Started

The Agent Development Kit (ADK) for .NET provides foundational components to build, test, and deploy intelligent agents.

Prerequisites

• .NET SDK 10.0+
• API Key (e.g., Google AI Studio, Vertex AI)

Authentication

Set the required environment variables in a .env file at the root of your project. The ADK automatically loads them via AdkEnv.Load().

Google AI Studio

GOOGLE_API_KEY=your_api_key_here

Vertex AI

GOOGLE_GENAI_USE_VERTEXAI=True
GOOGLE_CLOUD_PROJECT=your_project_id
GOOGLE_CLOUD_LOCATION=us-central1

First Agent in 30 Seconds

1. Create a new console app and install the package:

   dotnet new console -n QuickstartAgent
   cd QuickstartAgent
   dotnet add package GoogleAdk --prerelease
   

2. Update Program.cs:

   using GoogleAdk.Core;
   using GoogleAdk.Core.Agents;
   using GoogleAdk.Core.Runner;
   
   AdkEnv.Load();
   
   var agent = new LlmAgent(new LlmAgentConfig
   {
       Name = "quickstart",
       Model = "gemini-2.5-flash",
       Instruction = "You are a concise, helpful assistant."
   });
   
   await ConsoleRunner.RunAsync(agent);
   

Running the Web UI

To test and debug your agents visually, you can use the built-in server:

await AdkServer.RunAsync(agent);

Learn more about ADK Web

Custom Runners

For deep integration into production apps, background services, or Discord bots, you can bypass the pre-built runners and construct the core Runner directly:

var runner = new InMemoryRunner("custom-app", agent);

await foreach (var evt in runner.RunAsync("user-1", "session-id", content))
{
    // Handle your own streams and events
}

Learn more about Custom Runners