Strands Agents

Tagline: "Build production-ready, multi-agent AI systems in a few lines of code"

Strands Agents is an opensource AI agents framework developed by AWS. AWS says, "Strands Agents is a simple-to-use, code-first framework for building agents."

Introduction to this tutorial

This is a hands-on tutorial introduction to Strands Agents. While I have used Strands as the vehicle for this training, the concepts will carry to other AI agents frameworks. Further, Strands is opensource so it is easy to adopt in your software ecosystem. While the defaults in Strands are set for use with AWS, but it easily supports several other environments as well.

I have tried to keep a bottom-up approach here, to make it a practitioner's guide to learn about agents. So, we will learn about AI agents by using the Strands package, exploring the features it offers, and based on that we will learn about what agents are, as opposed to first learning the definition of agents and then see what this software has to match that. That is why I have organized it as a hands-on tutorial.

In this tutorial we have used a lot of content from Strands Agents documentation, and organized it in a way that makes it easy to follow and try out on your own computer.

The focus of this tutorial is on getting you started with the information which I believe you will need to develop your first several PoCs. In other words, certain advanced aspects have been skipped, as it is unlikely you will need them initially. When you have gone through this tutorial, it is highly recommended you go through the mentioned Stands Agents documentation to learn about the more involved aspects.

Note: While we are specifically discussing about AI agents ('agents' is a more generic term), but in this material "agents" and "AI agents" are used interchangeably.

Preparing to use this tutorial

• Place all the provided files in a folder - files are in the provided .zip.
• I suggest you a create conda environment for this tutorial and install the required dependencies in it.
  • The tutorial has been tested on Python 3.11.11
  • To install the required packages, please use pip install -r requirements.txt
• Since Strands has been developed by AWS, it uses Amazon Bedrock as the default for language models, but does support several other model providers. In this tutorial we will use OpenAI, so please get an OpenAI API key and place it in the .env file.
• I used VS Code to run these notebooks.

Happy learning!