What model providers does the n8n node support?

The node works with Amazon Bedrock, OpenAI, Google Gemini, and LiteLLM-supported providers, and you can switch providers between turns of the same conversation.

How does the node handle memory and user isolation?

Memory is managed automatically and scoped by Actor ID and Session ID. Each user can have an Actor ID (such as user-alice), and within that actor, multiple sessions are isolated from each other. Different actors with the same session ID get their own separate memory.

Can agents run privately in a VPC?

Yes. You set VPC mode on the Amazon Bedrock AgentCore API credential by entering your VPC Subnet IDs and Security Group IDs, and every agent that credential provisions runs privately. The harness pulls its managed container image from a private Amazon ECR repository in the same Region, so you need VPC endpoints for Amazon ECR and Amazon S3 rather than a NAT gateway.

What tools can agents access?

Agents can use the AgentCore Code Interpreter (for running code in a sandboxed environment), a cloud browser, AgentCore Gateway, and remote Model Context Protocol servers. Skills—bundles of instructions and scripts—can be loaded from the AWS curated catalog, Git repositories, Amazon S3, or filesystem paths.

Large Language ModelsAI Coding AssistantsAmazon AI BlogAug 5, 2026Aug 6, 2026, 04:00 JST

n8n adds Amazon Bedrock AgentCore node for production AI agents

Send on LINE

3 Key Points

What happened
Amazon Bedrock AgentCore harness is now generally available with a new open-source community node for n8n that lets users build AI agents in the visual editor without writing infrastructure code. The node (@aws/n8n-nodes-agentcore, version 0.3) works with Amazon Bedrock, OpenAI, Google Gemini, and LiteLLM-supported providers, and supports features like persistent memory, code interpreter tools, and VPC deployment.
Why it matters
Production agents need more than a single model call—they require persistent memory across sessions, access to real tools like browsers and sandboxes, and the ability to handle longer tasks. The node automates the scaffolding layer that typically consumes most development time: orchestration loops, tool calls, context management, state persistence, and failure recovery. Teams can now build multi-turn agents with managed memory, isolated sessions, and tool access directly in n8n's editor without switching between platforms.
What to watch
The node creates and reuses agents automatically on first run if you leave the Harness ARN blank, or accepts an existing harness ARN to invoke agents created outside n8n. Agents can be scoped to individual users via Actor ID (keeping each user's memory separate), run in your own VPC using private Amazon ECR and S3 endpoints instead of a NAT gateway, and load skills from the AWS curated catalog, Git repositories, Amazon S3, or local filesystem paths. The node is open source under the MIT license and powered by Strands Agents, AWS's open source agent framework.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

The release of Amazon Bedrock AgentCore harness with n8n integration addresses a significant gap in the agent-building workflow. n8n's built-in AI Agent node supports single model calls, but production systems require orchestration logic, memory management, tool integration, and failure handling—layers that teams typically build themselves. By exposing the full AgentCore harness in n8n's visual editor, AWS enables non-infrastructure specialists to assemble agents that persist state, isolate user sessions, and execute real tasks (code, browsing, API calls) without leaving the editor or writing custom code.

The node's flexibility around model providers (Bedrock, OpenAI, Gemini, LiteLLM) and its support for switching models mid-conversation reflect the increasingly multi-vendor AI landscape. The Actor ID and Session ID hierarchy allows the same agent to serve many users with clean memory isolation—a critical requirement for SaaS and multi-tenant systems. The ability to run agents in a VPC using only ECR and S3 endpoints (rather than internet access) makes it practical for enterprises with strict network policies.

The integration is grounded in Strands Agents, AWS's open source agent framework, and the community node itself is published under the MIT license, signaling AWS's intent to embed agent capabilities into existing workflow tools rather than forcing adoption of a proprietary platform.

FAQ

What model providers does the n8n node support?: The node works with Amazon Bedrock, OpenAI, Google Gemini, and LiteLLM-supported providers, and you can switch providers between turns of the same conversation.
How does the node handle memory and user isolation?: Memory is managed automatically and scoped by Actor ID and Session ID. Each user can have an Actor ID (such as user-alice), and within that actor, multiple sessions are isolated from each other. Different actors with the same session ID get their own separate memory.
Can agents run privately in a VPC?: Yes. You set VPC mode on the Amazon Bedrock AgentCore API credential by entering your VPC Subnet IDs and Security Group IDs, and every agent that credential provisions runs privately. The harness pulls its managed container image from a private Amazon ECR repository in the same Region, so you need VPC endpoints for Amazon ECR and Amazon S3 rather than a NAT gateway.
What tools can agents access?: Agents can use the AgentCore Code Interpreter (for running code in a sandboxed environment), a cloud browser, AgentCore Gateway, and remote Model Context Protocol servers. Skills—bundles of instructions and scripts—can be loaded from the AWS curated catalog, Git repositories, Amazon S3, or filesystem paths.

Amazon AI BlogRead Original Article

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.