LangChain in n8n

n8n's AI nodes implement LangChain's JavaScript framework. Each node is configurable: choose your own agent, LLM, memory, and other components. You can connect any other n8n node to your LangChain nodes as normal, so you can combine LangChain logic with any other data source or service n8n supports.

This page is for readers who already know LangChain and want to see how its concepts map onto n8n. If you're new to AI concepts, start with Understand AI components instead.

How LangChain concepts map to n8n nodes

n8n represents most LangChain concepts as cluster nodes: a root node that defines the cluster's main functionality, with one or more sub-nodes attached to extend it.

LangChain concept	Node category	Example nodes	Learn more
Chain	Root node	Basic LLM Chain, Question and Answer Chain, Summarization Chain	What's a chain in AI?
Agent	Root node	AI Agent	What's an agent in AI?
Language model	Sub-node	Anthropic Chat Model, OpenAI Chat Model, Ollama Chat Model	Sub-nodes
Vector store	Root node	Pinecone Vector Store, Qdrant Vector Store, Simple Vector Store	Store and search data with vectors
Memory	Sub-node	Postgres Chat Memory, Redis Chat Memory, Simple Memory	How memory works
Tool	Sub-node	Call n8n Workflow Tool, Custom Code Tool, Wikipedia	What's a tool in AI?
Retriever	Sub-node	Vector Store Retriever, Workflow Retriever	Retrieve relevant context
Embeddings	Sub-node	Embeddings OpenAI, Embeddings Cohere	Retrieve relevant context
Document loader	Sub-node	Default Data Loader, GitHub Document Loader	Sub-nodes
Output parser	Sub-node	Structured Output Parser, Auto-fixing Output Parser	Sub-nodes
Text splitter	Sub-node	Recursive Character Text Splitter, Token Splitter	Sub-nodes

Memory sub-nodes only attach to the AI Agent root node. Unlike LangChain, none of n8n's chain nodes support memory, so they can't reference earlier messages in a conversation. If your workflow needs to do that, use an agent instead of a chain.

The examples above aren't exhaustive. For the full, up-to-date list of nodes in each category, browse the root nodes and sub-nodes libraries. n8n also provides the Chat Trigger node to start a workflow from a chat message, and the LangChain Code node, which lets you write LangChain JavaScript code directly for functionality that doesn't have a dedicated n8n node yet.

Use LangSmith with a self-hosted n8n instance

LangSmith is a developer platform from the LangChain team. Connect a self-hosted n8n instance to LangSmith to record and monitor n8n executions, the same way you'd trace a LangChain application.

Feature availability

LangSmith tracing is available on:

• Self-hosted: All editions

It isn't available on n8n Cloud.

To connect n8n to LangSmith:

1. Log in to LangSmith and get your API key.

2. Set the following environment variables in the environment where you host n8n:

   Variable	Value
   LANGCHAIN_ENDPOINT	https://api.smith.langchain.com
   LANGCHAIN_TRACING_V2	true
   LANGCHAIN_API_KEY	Your LangSmith API key
   LANGCHAIN_PROJECT	Optional project name. Defaults to default
   LANGCHAIN_CALLBACKS_BACKGROUND	true to upload traces asynchronously

3. Restart n8n.

LangSmith creates a project named default after n8n sends its first trace, unless you set LANGCHAIN_PROJECT to a different name. Every trace then goes to that project instead. Traces can arrive with a short delay, because LANGCHAIN_CALLBACKS_BACKGROUND uploads them asynchronously by default. Set it to false for synchronous uploads when you're debugging.

For information on using LangSmith itself, refer to LangSmith's documentation.

Learn more about LangChain

You don't need to know LangChain to use n8n, but these resources can help if you want to go deeper:

• The LangChain documentation covers key concepts and use cases. Choose LangChain | Python or LangChain | JavaScript for quickstarts, code examples, and API documentation. LangChain also provides code templates for Python and JavaScript, with ideas for common patterns and use cases.

• What's LangChain? explains LangChain terminology in plain language, for a general audience.

• This YouTube series by Greg Kamradt works through the LangChain documentation with code examples.

• Join the n8n Discord to discuss LangChain and share your projects with the n8n community.