1. Getting Started

This blog is part of the ADK Masterclass - Hands-On Series. We'll explore the fundamentals of Google's Agent Development Kit (ADK) and build our first AI agent.

What is ADK?

• Model-agnostic: Can plug in any model like OpenAI, Ollama, Llama 3, etc. ADK supports all major LLM providers, making it versatile for different use cases.
• Deployment-agnostic: ADK can be deployed on any infrastructure, from local development to production, on our laptop, in a Docker container, on Cloud Run, on Vertex AI Agent Engine, GKE, or on any custom infrastructure.
• Compatible with other frameworks: Can work together with other libraries and tools, such as LangChain, CrewAI, OpenAI function calling, FastAPI, or any Python ecosystem.

Key Features

• Flexible Orchestration: Build workflows using Sequential, Parallel, or Loop steps, or let the LLM choose what to do next.
• Multi-Agent Architecture: Combine multiple specialized agents to create a modular, scalable system.
• Rich Tool Ecosystem: Add built-in tools, custom functions, third-party libraries, or even other agents as tools.
• Deployment Ready: Run agents locally, in containers, on Vertex AI Agent Engine, or any custom environment.
• Built-in Evaluation: Evaluate both the final output and each step to understand and improve agent performance.

Prerequisites

Before getting started, make sure we have:

• Python 3.11 or higher installed
• ADK version 1.18.0 or higher (required for Visual Agent Builder features)
• Google Cloud project

Setting Up Our Environment

1. Get API Keys

To use ADK with Google's AI models, we'll need an API key. Get one from Google AI Studio:

1. Visit https://aistudio.google.com/api-keys
2. Sign in with your Google account
3. Click "Create API Key"
4. Copy the generated API key

Set the API key as an environment variable:

export GOOGLE_API_KEY="our-api-key-here"

2. Set Up Our Project

Create a virtual environment:

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Installation

Install ADK using pip:

pip install google-adk

Verify the installation by checking the version:

❯ adk --version
adk, version 1.18.0

We can also view all available commands:

❯ adk --help

Usage: adk [OPTIONS] COMMAND [ARGS]...

  Agent Development Kit CLI tools.

Options:
  --version  Show the version and exit.
  --help     Show this message and exit.

Commands:
  api_server   Starts a FastAPI server for agents.
  conformance  Conformance testing tools for ADK.
  create       Creates a new app in the current folder with prepopulated agent template.
  deploy       Deploys agent to hosted environments.
  eval         Evaluates an agent given the eval sets.
  eval_set     Manage Eval Sets.
  run          Runs an interactive CLI for a certain agent.
  web          Starts a FastAPI server with Web UI for agents.

4. Building Our First Agent

Let's create our first agent using the ADK CLI. Run the following command:

adk create simple_agent

The CLI will guide us through the setup process. We'll be prompted to:

1. Choose a model for the root agent:
   • 1. gemini-2.5-flash
   • 2. Other models (fill later)
   Select option 1 for gemini-2.5-flash.
2. Choose a backend:
   • 1. Google AI
   • 2. Vertex AI
   Select option 1 for Google AI.
3. Enter Google API key: If we don't have an API key, we can create one at https://aistudio.google.com/apikey. Enter our API key when prompted.

Here's what the interactive session looks like:

❯ adk create simple_agent

Choose a model for the root agent:

1. gemini-2.5-flash

2. Other models (fill later)

Choose model (1, 2): 1

1. Google AI

2. Vertex AI

Choose a backend (1, 2): 1

Don't have API Key? Create one in AI Studio: https://aistudio.google.com/apikey

Enter Google API key [YOUR_API_KEY]:

Agent created in /Users/arjunprabhulal/google-adk-masterclass/1-adk-getting-started/simple_agent:

- .env
- __init__.py
- agent.py

Our agent project has been created with the following structure:

└── simple_agent
    ├── __init__.py
    ├── .env
    └── agent.py

2 directories, 3 files

The .env file contains our configuration:

cat simple_agent/.env

GOOGLE_GENAI_USE_VERTEXAI=0
GOOGLE_API_KEY=YOUR_API_KEY

The agent.py file contains our agent definition:

from google.adk.agents import Agent

root_agent = Agent(
    model='gemini-2.5-flash',
    name='root_agent',
    description='A helpful assistant for user questions.',
    instruction='Answer user questions to the best of your knowledge',
)

Next Steps

Now that we've created our first agent, we're ready to dive into more topics:

• Learn about Setting Up Agents using different methods (Web, CLI, Programmatic)
• Explore the full ADK Masterclass for comprehensive learning

Resources

• Official ADK Documentation
• ADK GitHub Repository