Build a streaming agent with Go¶

With this quickstart, you'll learn to create a simple agent and use ADK Streaming to enable voice and video communication with it that is low-latency and bidirectional. We will install ADK, set up a basic "Google Search" agent, and run a local server to interact with it using the embedded ADK Web UI.

Note: This guide assumes you have experience using a terminal in Linux, Mac, or Windows environments.

Supported models for voice/video streaming¶

In order to use voice/video streaming in ADK, you will need to use Gemini models that support the Live API.

You can find the complete model ID(s) that support the Gemini Live API in the documentation:

1. Set up the platform¶

Choose a platform from either Google AI Studio or Google Cloud Agent Platform (Vertex AI):

Option 1: Google AI Studio¶

Get an API key from Google AI Studio.
Set the environment variables in your terminal:

export GOOGLE_GENAI_USE_VERTEXAI=FALSE
export GOOGLE_API_KEY=PASTE_YOUR_ACTUAL_API_KEY_HERE

Option 2: Vertex AI / Agent Platform¶

Set up a Google Cloud project.
Set up the gcloud CLI and authenticate via terminal:

gcloud auth login

Find your active Google Cloud Project ID with:

gcloud config get-value project

Enable the Agent Platform API.
Set the environment variables in your terminal, updating your project ID and location:

export GOOGLE_GENAI_USE_VERTEXAI=TRUE
export GOOGLE_CLOUD_PROJECT=PASTE_YOUR_ACTUAL_PROJECT_ID
export GOOGLE_CLOUD_LOCATION=us-central1

2. Setup Environment & Install ADK¶

Create a new directory for your project and initialize a Go module:

mkdir adk-streaming
cd adk-streaming
go mod init adk-streaming

Install the ADK Go SDK:

go get google.golang.org/adk@latest

3. Project Structure¶

Create a single main.go file. You do not need to copy or host static assets manually since the ADK launcher packages the Web UI for you.

adk-streaming/
└── main.go

4. Create the Agent and Launcher¶

Copy-paste the following code into your main.go file.

package main

import (
    "context"
    "log"
    "os"

    "google.golang.org/genai"

    "google.golang.org/adk/agent"
    "google.golang.org/adk/agent/llmagent"
    "google.golang.org/adk/cmd/launcher"
    "google.golang.org/adk/cmd/launcher/full"
    "google.golang.org/adk/model/gemini"
    "google.golang.org/adk/tool"
    "google.golang.org/adk/tool/geminitool"
)

func main() {
    ctx := context.Background()

    // Initialize the Gemini model
    // Configured to dynamically support Google AI Studio or Vertex AI backends
    var cfg *genai.ClientConfig
    if apiKey := os.Getenv("GOOGLE_API_KEY"); apiKey != "" {
        cfg = &genai.ClientConfig{APIKey: apiKey}
    }

    // Set model ID dynamically based on the platform option chosen:
    // - For Google AI Studio (Gemini API), use "gemini-2.5-flash-native-audio-preview-12-2025"
    // - For Vertex AI (Agent Platform), use "gemini-live-2.5-flash-native-audio"
    modelID := "gemini-2.5-flash-native-audio-preview-12-2025"
    if os.Getenv("GOOGLE_GENAI_USE_VERTEXAI") == "TRUE" {
        modelID = "gemini-live-2.5-flash-native-audio"
    }
    model, err := gemini.NewModel(ctx, modelID, cfg)
    if err != nil {
        log.Fatalf("Failed to create model: %v", err)
    }

    // Define the agent
    a, err := llmagent.New(llmagent.Config{
        Name:        "bidi-demo",
        Model:       model,
        Description: "Agent optimized for real-time bidirectional streaming.",
        Instruction: "You are a helpful voice assistant. Answer questions concisely.",
        Tools: []tool.Tool{
            geminitool.GoogleSearch{},
        },
    })
    if err != nil {
        log.Fatalf("Failed to create agent: %v", err)
    }

    // Configure the launcher with the agent loader
    config := &launcher.Config{
        AgentLoader: agent.NewSingleLoader(a),
    }

    // Create a full launcher with built-in web and console capabilities
    l := full.NewLauncher()
    if err = l.Execute(ctx, config, os.Args[1:]); err != nil {
        log.Fatalf("Run failed: %v\n\n%s", err, l.CommandLineSyntax())
    }
}

5. Try the agent¶

Run the application with the web, api, and webui arguments. This boots up the web server, registers the REST API backend, and serves the ADK Web UI dashboard simultaneously:

go run main.go web api webui

Once the server starts, open the Web UI in your browser:

http://localhost:8080/ui/

Try with voice and video¶

Click the microphone button to enable voice input, and ask questions. The agent will respond in real-time via audio. If you have a camera enabled, you can also stream video and ask "What do you see?".

Caveat¶

You cannot use text chat with the native-audio models if the model is configured for audio-only input/output in the connection. However, the live API supports both.
Ensure you use earphones to prevent echo when using voice input/output.

Congratulations! You've successfully created and interacted with your first Streaming agent using ADK in Go!