Amazon Quick, a platform for building generative AI applications, now allows integration with AgentCore Runtime for hosting Model Context Protocol (MCP) servers. This integration enables AI agents to access external data and tools securely, supporting stateful, multi-turn interactions. The solution is designed to promote reusability and reduce duplication of AI tools, allowing customers to use products within Amazon Quick without custom connectors for each use case. Source: awsml
The integration is handled through connectors on the Amazon Quick end and AgentCore Gateway on the AgentCore end. AgentCore Gateway and Runtime are part of Amazon Bedrock AgentCore, a fully managed service for generative AI applications. The authorization flow from Amazon Quick to AgentCore Gateway is called Inbound Auth, which uses Amazon Cognito for authorization, though other identity providers can be used. Outbound Auth handles machine-to-machine authentication and authorization using AgentCore Identity, which is tailored for AI agents. The MCP protocol currently requires OAuth 2.0 as the authentication protocol, so Outbound Auth uses OAuth 2.0. Source: awsml
Amazon Quick can be used in a web browser or desktop app to work with chat agents or Flows that provide AI agent capabilities. To connect the AI agent with the MCP server for access to additional tools and sub-agent capabilities, the MCP server must be integrated with Amazon Quick. This integration involves deploying a sample MCP server on AgentCore Runtime, configuring it with the MCP protocol, and setting up authentication flows for both inbound and outbound connections. The solution also includes steps for testing and cleaning up resources after deployment. Source: awsml
