How we built an MCP bridge to give our AgentCore-hosted AI agent access to local MCP tools
AWS engineers built a WebSocket/native-messaging bridge letting cloud agents call local MCP servers.
“This is the same pattern that powers products like Claude Cowork, a cloud agent calling local tools through MCP, but fully self-hosted on AWS with your own model and custom tool servers.”
AWS published a detailed engineering walkthrough for bridging cloud-hosted AI agents (on Amazon Bedrock AgentCore) to locally-running MCP servers via WebSocket tunneling through a browser extension. The pattern solves a real gap in the MCP spec — remote client, local server — and mirrors the architecture Anthropic uses for Claude Cowork. An internal finance assistant built on this pattern logged over 41,000 conversations in its first year, signaling production viability.