playground-agentic-ai

Example of how to create agents, MCP servers, and clients - all of that with Strands Agents (in Python 🐍), Spring AI, FastMCP, and deploy that with FastAPI via containers, on AWS Lambda via AWS SAM, and Amazon Bedrock AgentCore.

Local Development

Here is a list of the recommended prerequisites for this repository.

• Pre-installed tools:
  • Finch or any other tool for local container development, that is compatible with Docker APIs.
  • Most recent AWS CLI (2.31.4 or higher).
  • Most recent AWS SAM CLI (1.145.2 or higher).
  • Node.js in version 22.20.x or higher.
  • Python in version 3.12.x or higher.
  • Java in version 21.x or higher.
• Configured profile in the installed AWS CLI with credentials for your AWS IAM user account of choice.

If you would like to start all the dependent services, run the following commands:

# After cloning it, inside the repository root:

$ finch vm start             # ... if `finch` did not start the virtual machine yet.

Resources

Strands Agents

• Building Intelligent Agents with Strands: A Hands-On Guide

AWS MCP

• AWS MCP Servers

Amazon Bedrock AgentCore

• Deploy ANY AI Agent to Production in Minutes - Amazon Bedrock AgentCore Tutorial
• AWS Workshop: Diving Deep into Amazon Bedrock AgentCore
• Spring AI: agent deployed locally and on AgentCore
• Spring AI: an MCP server example deployed locally and on AgentCore

License

• MIT

Authors

• Wojciech Gawroński (afronski) (aka AWS Maniac)