n8n Workflows Now Run Production AI Agents with Amazon Bedrock AgentCore Harness

The AWS ML blog announced that Amazon Bedrock AgentCore harness, a managed capability for building and running production AI agents, is now generally available and integrated into n8n via a new open-source community node, @aws/n8n-nodes-agentcore. This node brings the full harness into n8n's visual editor, offering agents with persistent memory, tools like a code interpreter and cloud browser, and the ability to use various model providers including Amazon Bedrock, OpenAI, Google Gemini, and LiteLLM-supported ones, with the option to switch providers between turns of the same conversation. The harness runs the orchestration loop, manages the context window, keeps state, recovers from failures, and isolates sessions, with each session in its own isolated environment. Users can configure an agent by leaving the Harness ARN blank for automatic provisioning or using an existing ARN. The node works with n8n Cloud or self-hosted instances, uses standard AWS credentials, and supports features like scoping memory per user with an Actor ID, adding skills from catalogs or Git, and running in a private VPC. It is open source under the MIT license and powered by Strands Agents.