One-Line Summary: Create the Python project, install Google ADK, and configure your Gemini API key.

Prerequisites: Python 3.12+ installed, a Google account

Create the Project

Open your terminal and set up the project directory:

# Create the project directory and navigate into it
mkdir research-agent
cd research-agent
 
# Create a virtual environment
python -m venv .venv
 
# Activate it (Linux/macOS)
source .venv/bin/activate
 
# Activate it (Windows)
# .venv\Scripts\activate

Install Google ADK

# Install the Agent Development Kit
pip install google-adk

That single package brings in everything you need:

What You GetPurpose
google-adkAgent framework — agents, tools, runner, sessions
google-genaiGemini API client (installed as a dependency)
adk CLICreate, run, test, and deploy agents from the terminal

Save the dependencies:

pip freeze > requirements.txt

Get Your Gemini API Key

  1. Go to Google AI Studio
  2. Click Create API Key
  3. Copy the key

Create a .env file in your project root:

# .env
GOOGLE_API_KEY=your-gemini-api-key-here

Cost: Gemini has a generous free tier. The gemini-2.5-flash model we will use is fast and affordable for development.

Create the Agent Package

ADK expects your agent to live in a Python package. Create the directory structure:

# Create the agent package directory
mkdir research_agent
 
# Create the required files
touch research_agent/__init__.py
touch research_agent/agent.py

Add the package init that tells ADK where to find your agent:

# research_agent/__init__.py
# ==========================================
# Package init — ADK discovers agents through this file
# ==========================================
 
from . import agent

Verify ADK Is Installed

Run a quick check to make sure everything is working:

# Check ADK version
adk --version

You should see a version number like 1.x.x. If you see a command not found error, make sure your virtual environment is activated.

Verify Your API Key

Create a quick test to confirm your Gemini API key works:

# test_setup.py
# ==========================================
# Quick smoke test — verifies API key works
# ==========================================
 
import google.genai as genai
import os
from dotenv import load_dotenv
 
load_dotenv()
 
client = genai.Client()
response = client.models.generate_content(
    model="gemini-2.5-flash",
    contents="Say hello in exactly 5 words.",
)
 
print(response.text)
print("\nSetup verified — you're ready to build.")

Run it:

python test_setup.py

If you see a five-word greeting and the success message, your environment is configured correctly. If you get an authentication error, double-check your API key in .env.

Your Project So Far

research-agent/
├── research_agent/
│   ├── __init__.py       ✅ Package init
│   └── agent.py          📝 Empty (we fill this next)
├── .env                  ✅ API key configured
├── requirements.txt      ✅ Dependencies saved
└── test_setup.py         ✅ Setup verified

Reference: ADK Python Get Started · Google AI Studio

← What We're Building | Next: Step 3 - Your First Agent →