Environment Setup

Get your development environment ready for AutoGen

Prerequisites: Python 3.10 or later is required for all AutoGen packages.

Choose Your AutoGen Framework

AutoGen offers different frameworks depending on your needs:

AutoGen Studio

Web-based UI for prototyping with agents without writing code.

Best for: Beginners, no-code prototyping

pip install -U autogenstudio

AgentChat

Programming framework for conversational single and multi-agent applications.

Best for: Python developers, conversational AI

pip install -U "autogen-agentchat"

AutoGen Core

Event-driven framework for building scalable multi-agent AI systems.

Best for: Advanced developers, scalable systems

pip install "autogen-core"

Step-by-Step Installation Guide

Step 0: Create AutoGen Root Folder (Recommended)

Create a dedicated folder to organize all your AutoGen projects, demos, and applications.

Recommended folder structure:


mkdir -p ~/Development/AutoGen

cd ~/Development/AutoGen



# Create subfolders for organization

mkdir demos applications tutorials experiments

This creates: ~/Development/AutoGen/ with organized subfolders

Recommended folder structure:


md C:\Development\AutoGen

cd C:\Development\AutoGen



# Create subfolders for organization

md demos applications tutorials experiments

This creates: C:\Development\AutoGen\ with organized subfolders

Folder Purpose:

demos/ - Tutorial examples
applications/ - Your projects
tutorials/ - Learning exercises
experiments/ - Testing new ideas

Benefits:

Organized project structure
Easy to find your work
Separates demos from real projects
Consistent development environment

Step 1: Create Virtual Environment (Recommended)

We recommend using a virtual environment to isolate AutoGen dependencies from your system. Navigate to your AutoGen root folder first:

Important: Run these commands from your AutoGen root folder created in Step 0

Linux/Mac:

python3 -m venv .venv
source .venv/bin/activate

Windows:

python -m venv .venv
.venv\Scripts\activate.bat

To deactivate later:

deactivate

conda create -n autogen python=3.10
conda activate autogen

Step 2: Install AutoGen Package

Choose the installation command based on your selected framework:

For AgentChat (Recommended for beginners):

pip install -U "autogen-agentchat"

For AutoGen Core:

pip install "autogen-core"

Step 3: Install Model Extensions

To use AI models, install the appropriate extensions:

For OpenAI models:

pip install "autogen-ext[openai]"

For Azure OpenAI:

pip install "autogen-ext[azure]"

API Keys Required: You'll need to set up API keys for OpenAI or Azure OpenAI to use these models.

Step 4: Install Docker (Optional)

Docker is recommended for secure code execution using DockerCommandLineCodeExecutor:

Follow the Docker installation guide for your operating system.

Verify Your Installation

Test Your Setup

Create a simple test file to verify your AutoGen installation:

Prerequisites: You'll need an OpenAI API key. Get one from OpenAI Platform

Step 1: Set Up API Key

Set your OpenAI API key as an environment variable:

Mac/Linux Instructions:

Important: Run this command in the same terminal where you'll run the test

1. Set the environment variable:

export OPENAI_API_KEY="sk-your-actual-api-key-here"

2. Verify it's set:

echo $OPENAI_API_KEY

Tip: To make this permanent, add the export command to your ~/.zshrc or ~/.bash_profile

Windows Instructions:

Option 1: Command Prompt (cmd)

1. Set the variable:

set OPENAI_API_KEY=sk-your-actual-api-key-here

2. Verify it's set:

echo %OPENAI_API_KEY%

Option 2: PowerShell

1. Set the variable:

$env:OPENAI_API_KEY="sk-your-actual-api-key-here"

2. Verify it's set:

echo $env:OPENAI_API_KEY

Tip: To make this permanent, add it through System Properties → Environment Variables

Step 2: Create test file: Navigate to your demos/ folder and create: test_installation.py

Mac/Linux:

cd ~/Development/AutoGen/demos
touch test_installation.py

Windows:

cd C:\Development\AutoGen\demos
type nul > test_installation.py

Step 3: Add the following code to test_installation.py:


import asyncio

from autogen_agentchat.agents import AssistantAgent

from autogen_ext.models.openai import OpenAIChatCompletionClient



async def main() -> None:

    agent = AssistantAgent("assistant", OpenAIChatCompletionClient(model="gpt-4o"))

    result = await agent.run(task="Say 'Hello World!'")

    print(result)



asyncio.run(main())

Step 4: Run the test:

python test_installation.py

What to Expect: The output will show AutoGen's internal message structure including your request, the agent's response ("Hello World!"), and metadata like token usage. This detailed output confirms that AutoGen is communicating properly with the OpenAI API and processing messages correctly.

Success Indicators:

No error messages
Message structure displayed
Agent responds with "Hello World!"
Token usage shown

Common Issues:

API key not set → Follow Step 1
Import errors → Check installation
Network issues → Check internet
Rate limits → Wait and retry

Success! If you see the success message, your AutoGen environment is ready!

🎉 Environment Ready!

Your AutoGen development environment is now set up. Ready to start learning?

Start Learning Fundamentals Try Interactive Tutorials