n8n AI Agent Workflow: JavaScript Custom Tool (LangChain)

Build a JavaScript Custom Tool AI Agent in n8n using LangChain

This n8n AI Agent Workflow lets you run an AI agent with a JavaScript Custom Tool (LangChain). It uses an OpenAI-backed chat trigger to receive the user’s request, then executes your custom JavaScript tool logic through the LangChain agent to produce the final response—fully inside n8n automation.

What this workflow does

• Starts with a Manual Trigger (so you can run and test the workflow on demand).
• Receives input via the n8nn8n-nodes-langchainchat trigger, routing the message into a LangChain agent flow.
• Initializes required values using the set node before agent execution.
• Runs an n8nn8n-nodes-langchainagent that can decide when to call tools.
• Implements your JavaScript Custom Tool using the n8nn8n-nodes-langchaintool code node, where the tool’s behavior is defined in code.
• Connects the agent to an OpenAI language model through the n8nn8n-nodes-langchainlm chat open ai node.

Use cases

• Extend an AI agent with bespoke business logic (e.g., custom parsing, formatting, or rule-based actions) using a JavaScript tool.
• Create internal automation copilots in n8n that can interpret requests and invoke your custom tool when needed.
• Prototype and test LangChain tool behaviors quickly with manual runs before integrating into larger production workflows.
• SaaS operations workflows where the agent must apply deterministic transformations via code, then respond with an AI-generated result.

Technical details

• Manual Trigger
• n8nn8n-nodes-langchainchat trigger
• set (variable initialization)
• n8nn8n-nodes-langchainagent (agent orchestration)
• n8nn8n-nodes-langchaintool code (JavaScript Custom Tool logic)
• n8nn8n-nodes-langchainlm chat open ai (OpenAI model for the chat agent)