According to Amazon AWS AI, the company has introduced stateful Model Context Protocol (MCP) client capabilities on Amazon Bedrock AgentCore Runtime. The new features enable developers to build more interactive and dynamic AI agent systems.
The announcement details three key capabilities now available for stateful MCP servers. According to the source, these include the ability to request user input during execution, invoke LLM sampling for dynamic content generation, and stream progress updates for long-running tasks. The post indicates that developers can access code examples demonstrating each of these capabilities.
The introduction of these stateful features represents an expansion of Amazon Bedrock’s agent development toolkit. According to the AWS announcement, the post provides guidance on deploying a working stateful MCP server, suggesting practical implementation is now available to developers working with the platform. The stateful capabilities appear designed to support more complex agent workflows that require ongoing interaction and real-time feedback during execution.