Introduction
In part 5, we looked at how to deploy and run our MCP client on AgentCore Runtime. Starting from this article, we'll look at the Spring AI AgentCore functionality. In this article, we'll use Spring AI AgentCore Starter in our sample application to deploy and run our MCP client on AgentCore Runtime. As Spring AI 2.0 and Spring Boot 4.1 have both already gone GA, I updated our application to use them; see pom.xml. For this, I created a new spring-ai-2.0-ac-conference-app-agent-bedrock-agentcore-runtime. It consists of the agent and Infrastructure as Code (IaC) subfolders.
Differences in the implementation of the MCP client for the Conference application between Spring AI 1.1 and 2.0
I have noticed 2 major differences in the implementation of the MCP client for the Conference application. We developed it mainly in part 3, and then adjusted it in part 4 and part 5. See SpringAIAgentController.java for the full version of the controller.
The first one is that ChatClient now requires passing the ChatOptions.Builder object to the defaultOptions method instead of the ChatOptions object itself:
