Microsoft AutoGen Studio 2.0 Local Setup Guide

In March 2021, Microsoft released AutoGen, a framework enabling the development of LLM applications using multiple agents that can collaborate to solve tasks.

While the technology was impressive, the setup process proved complex for the average user, leaving many, including myself, hesitant to attempt installation on our machines.

Today, Microsoft unveiled a new application that simplifies the process of prototyping AI agents—it's called AutoGen Studio.

What is AutoGen Studio?

AutoGen Studio is a user interface application that is built on top of the AutoGen framework, which is a toolkit for building AI agents.

According to Microsoft, AutoGen Studio is designed to facilitate the rapid prototyping of multi-agent workflows and demonstrate end-user interfaces. Take note, though, that it is not yet intended as a production-ready application.

Here’s an example of AutoGen Studio in action:

Prompt: Plot a chart of NVDA and TESLA stock price YTD. Save the result to a file named nvda_tesla.png

AutoGen combines multiple AI agents to generate code for retrieving data from the web and creating an image with stock prices.

That’s pretty cool, right?

In this article, I will guide you through the step-by-step process of installing and running AutoGen Studio on your local machine.

Let’s get started.

Step #1: Install Python

AutoGen Studio 2.0 requires Python 3.11. Ensure you have this version installed to avoid compatibility issues.

You can check your Python version by running this command in your terminal:

Python --version

In March 2021, Microsoft released AutoGen, a framework enabling the development of LLM applications using multiple agents that can collaborate to solve tasks.

While the technology was impressive, the setup process proved complex for the average user, leaving many, including myself, hesitant to attempt installation on our machines.

Today, Microsoft unveiled a new application that simplifies the process of prototyping AI agents—it's called AutoGen Studio.

What is AutoGen Studio?

AutoGen Studio is a user interface application that is built on top of the AutoGen framework, which is a toolkit for building AI agents.

According to Microsoft, AutoGen Studio is designed to facilitate the rapid prototyping of multi-agent workflows and demonstrate end-user interfaces. Take note, though, that it is not yet intended as a production-ready application.

Here’s an example of AutoGen Studio in action:

Prompt: Plot a chart of NVDA and TESLA stock price YTD. Save the result to a file named nvda_tesla.png

If you don’t have it installed on your machine yet, go to this link to download and install Python 3.11.

Step #2: Install Conda

Conda is an open-source package and environment management system that operates on various operating systems. It helps AutoGen manage dependencies effectively without affecting other projects or system-wide settings.

Download and install Conda from here.

To verify the installation, check the Conda version on your system using this command:

conda — version

Step #3: Get an OpenAI Key

AutoGen will utilize large language models like OpenAI’s GPT-4 to execute tasks. To use these models, you need an OpenAI key.

Create a new key by visiting your OpenAI account, navigating to the API keys tab, and generating a new key by clicking the “Create new secret key” button.

Step #4: Setup your environment

The first step is creating a virtual environment for AutoGen Studio. To do this, use Conda by executing the following command in a terminal:

conda create -n autogenstudio python=3.11

When you get prompted to install the following packages, select “yes” to continue:

Activate the AutoGen Studio environment with this command:

conda activate autogenstudio

You should see the following interface in your terminal:

Next, set up the OpenAI secret key using this command:

set OPENAI_API_KEY=your_openai_api_key

And that’s it. Your environment setup is complete.

Step #5: Install AutoGen Studio

In this step, we will install AutoGen Studio 2.0. In the same Conda terminal from previous steps, run the command:

pip install autogenstudio

This step will take a few minutes as it downloads and installs AutoGen Studio along with its dependencies. Feel free to grab a snack or coffee while you wait.

Once it’s done, launch AutoGen Studio with this command:

autogenstudio ui --port 8081

The app will open in your local browser, and you should see the following interface:

If the app doesn’t launch automatically, open a web browser and go to this URL: http://127.0.0.1:8081/

Congratulations! You have successfully set up AutoGen Studio 2.0. Now you can begin creating AI agents.

To verify that the setup is working correctly, I created a simple AI agent by clicking the “New” button and selecting a visualization agent workflow. Here’s an example prompt:

Plot a chart of NVDA and TESLA stock price for 2023. Save the result to a file named nvda_tesla.png

The agent successfully generated the requested CSV files.

One thing to note is the absence of support for streamed responses in the user interface for real-time progress monitoring. The results are displayed all at once when the agents complete their tasks.

To monitor progress, you can open the terminal and observe the exchange of messages between AI agents in real-time.

I hope the engineering team can enhance the user experience in the future, but even in its current state, AutoGen Studio is a promising tool.

In the next series of articles, I will delve deeper into how to use AutoGen Studio’s various features to create and manage AI agents and workflows.

Final Thoughts

In conclusion, the release of AutoGen Studio by Microsoft is a significant step forward in simplifying the process of building AI agents. While the initial setup steps may still require some technical know-how, it’s heartening to see efforts being made to make AI development more accessible to the average user.

I hope you find this article interesting, and if you followed the guide, I hope you were able to setup the application on your local machine. If you have any questions or encountered any challenges while setting up AutoGen Studio, please feel free to share them in the comments, and I’ll be happy to assist you.