n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Advanced AI Table of contents

Tutorial: Build an AI workflow Related resources
in n8n
Node types
Build AI functionality using n8n: from creating your own chat bot, to using AI to
LangChain in n8n Workflow templates
process documents and data from other sources.
Overview Chat trigger
Langchain concepts in n8n Chatbot widget
Feature availability
LangChain learning
resources This feature is available on Cloud and self-hosted n8n, in version 1.19.4 and above.

Examples and concepts

Introduction Get started Explore examples and concepts
What is a tool?
Work through the short tutorial to Browse examples and workflow
Use Google Sheets as a data
source
learn the basics of building AI templates to help you build. Includes
workflows in n8n. explanations of important AI
Call an API to fetch data
concepts.
Set a human fallback for AI Tutorial
workflows Examples

How n8n uses LangChain Browse AI templates

Learn more about how n8n builds on Explore a wide range of AI workflow
LangChain. templates on the n8n website.

LangChain in n8n AI workflows on n8n.io

Related resources
Related documentation and tools.

Node types

This feature uses Cluster nodes: groups of root and sub nodes that work together.

Cluster nodes are node groups that work together to provide functionality in an n8n
workflow. Instead of using a single node, you use a root node and one or more sub-
nodes. This allows extensive configuration of node functionality.

Workflow templates

You can browse workflow templates in-app or on the n8n website Workflows page.

Refer to Templates for information on accessing templates in-app.

Chat trigger

Use the n8n Chat Trigger to trigger a workflow based on chat interactions.

Chatbot widget

n8n provides a chatbot widget that you can use as a frontend for AI-powered chat
workflows. Refer to the @n8n/chat npm page for usage information.

Was this page helpful?

Previous Next
Pagination Tutorial: Build an AI workflow in n8n

Change cookie settings

Made with Material for MkDocs Insiders
n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Tutorial: Build an AI workflow in n8n Table of contents

Tutorial: Build an AI workflow Step one: Sign up for n8n
in n8n
Step two: Create a new
This tutorial introduces LangChain functionality in n8n. You can work through it with workflow
LangChain in n8n
no prior knowledge of LangChain, AI, or n8n. However, if you've never used n8n Step three: Add a trigger
Overview
before, you should also do the Longer quickstart, which introduces key n8n node
Langchain concepts in n8n
concepts. Step four: Add an agent
LangChain learning
Step five: Chat functionality
resources
In this tutorial you will:
Step six: Get data from
Examples and concepts another n8n workflow
Create a workflow from scratch. It uses the Chat Trigger to simulate and test
Introduction Step seven: Test the
chat interactions, ChatGPT to power the chat functionality, and a custom tool to workflow
What is a tool?
connect to other n8n workflows. Next steps
Use Google Sheets as a data
source Understand key concepts, including:
Call an API to fetch data The role of agents, models, and tools when building AI functionality.
Set a human fallback for AI
workflows Cluster nodes in n8n: groups of root and sub-node.

Loading your own data into AI workflows.

Step one: Sign up for n8n

Skip this section if you've already installed n8n or signed up for a Cloud account

n8n is available as a Cloud service, npm module, and Docker image. For this
quickstart, n8n recommends using Cloud. A free trial is available for new users.

n8n version 1.21.0 or above

This tutorial requires n8n version 1.21.0 or above.

Step two: Create a new workflow

When you open n8n, you'll see the Workflows list.

To create a new workflow for this tutorial:

If you have no workflows, select Start from scratch.

If you already built a workflow, select Add workflow.

Step three: Add a trigger node

A trigger node starts a workflow. For this tutorial, use the Chat Trigger. This provides
a chat interface that you can use to test the workflow.

1. Select Add first step.

2. Search for Chat Trigger. n8n shows a list of nodes that match the search.

3. Select Chat Trigger to add the node to the canvas. n8n opens the node.

4. Close the node details view to return to the canvas.

Step four: Add an agent

An agent takes an input, such as a user query, then uses a collection of tools to
respond. In n8n the AI Agent node is a root node: a node whose behavior you can
configure using sub-nodes.

Cluster nodes: root and sub-nodes

1. Select the Add node connector. n8n opens the nodes panel.

2. Search for Agent. n8n shows a list of nodes that match the search.

3. Select AI Agent. n8n adds the node to the canvas and opens it.

4. Close the node details view to return to the canvas.

Step five: Chat functionality

To provide chat interactions, the workflow needs:

A chat model, to process chat inputs and generate responses.

Memory, so that the workflow can remember and use previous sections of the
chat.

Use the connectors on the bottom of the AI Agent node to connect the following:

Model: OpenAI Chat Model. Set up the credentials for this node.

Credential setup steps

Memory: Window Buffer Memory

Step six: Get data from another n8n workflow

One unique feature of AI in n8n is the ability to pull in data from other n8n workflows
using the Custom n8n Workflow Tool node. This means you can:

Load your own custom data

Create custom functionality in another workflow

Because n8n can connect to any service with a public API, this is a powerful tool.

This example generates some fake data in a workflow, and loads it in to the AI
workflow. The AI workflow passes a parameter to the workflow generating the data,
to ensure it only returns data that's intended to be public.

1. Create a new workflow, then copy in this workflow JSON:

1 {
2 "name": "Workflow tool example",
3 "nodes": [
4 {
5 "parameters": {
6 "jsCode": "return [{\"fruit\": \"apple\", \"color\": \"green\", \"dataPrivacyLevel\": \"public\"},{\"fruit\": \"ba
7 },
8 "id": "4a30b74c-ed27-4668-8e68-c2e3a630ecd9",
9 "name": "Code",
10 "type": "n8n-nodes-base.code",
11 "typeVersion": 2,
12 "position": [
13 1180,
14 500
15 ]
16 },
17 {
18 "parameters": {
19 "conditions": {
20 "string": [
21 {
22 "value1": "={{ $json.dataPrivacyLevel }}",
23 "value2": "={{ $('Execute Workflow Trigger').item.json.visibility }}"
24 }
25 ]
26 }
27 },
28 "id": "eb479841-8099-4537-b89b-4b9867446688",
29 "name": "Filter",
30 "type": "n8n-nodes-base.filter",
31 "typeVersion": 1,
32 "position": [
33 1380,
34 500
35 ]
36 },
37 {
38 "parameters": {
39 "aggregate": "aggregateAllItemData",
40 "include": "specifiedFields",
41 "fieldsToInclude": "fruit, color",
42 "options": {}
43 },
44 "id": "6d6c847d-7417-4780-8a8b-b99996cd6166",
45 "name": "Aggregate",
46 "type": "n8n-nodes-base.aggregate",
47 "typeVersion": 1,
48 "position": [
49 1600,
50 500
51 ]
52 },
53 {
54 "parameters": {},
55 "id": "1ec54b09-7533-4280-8bf2-54407d6bf134",
56 "name": "Execute Workflow Trigger",
57 "type": "n8n-nodes-base.executeWorkflowTrigger",
58 "typeVersion": 1,
59 "position": [
60 1000,
61 500
62 ]
63 }
64 ],
65 "pinData": {},
66 "connections": {
67 "Code": {
68 "main": [
69 [
70 {
71 "node": "Filter",
72 "type": "main",
73 "index": 0
74 }
75 ]
76 ]
77 },
78 "Filter": {
79 "main": [
80 [
81 {
82 "node": "Aggregate",
83 "type": "main",
84 "index": 0
85 }
86 ]
87 ]
88 },
89 "Execute Workflow Trigger": {
90 "main": [
91 [
92 {
93 "node": "Code",
94 "type": "main",
95 "index": 0
96 }
97 ]
98 ]
99 }
100 }
101 }

The workflow uses the Code node to generate a sample data set, containing a
list of fruits, their colors, and whether this information should be public or
private.

2. Copy the workflow ID from workflow URL. The ID is the group of random
numbers and letters at the end of the URL.

3. In the AI workflow, select the Tool output on the AI Agent. n8n opens the nodes
panel.

4. Select Custom n8n Workflow Tool. n8n adds the node to the canvas and opens
it.

5. Configure the node parameters:

Name: Fruit

Description: Call this tool to get a list of fruit.

Workflow ID: Paste in the workflow ID that you copied from the example
workflow URL.

Response Property Name: data . This is the name of the item created in the
Aggregate node in the example workflow.

6. Select Add Value in Workflow Values to pass information to the workflow you're
calling. Configure the value as follows:

Name: visibility

Type: String

Value: public

n8n makes this value available to the workflow in the output data of the trigger
node in the workflow you're calling. In this example, to access the value in an
expression, use {{ $('Execute Workflow Trigger').item.json.visibility
}} . You can see this in practice in the Filter node in the sample workflow.

Step seven: Test the workflow

You can now try out the workflow: close any open nodes, then select Chat to start
talking with the AI. Ask Please give me a list of fruits . The AI Agent node
executes the workflow you specified in Workflow ID, passing in the value of
visibility from the Workflow Values field. The workflow generates some sample
data, filters it to remove any items that aren't set to public , and returns the
resulting data.

Next steps
Explore n8n using the Quickstarts.

Learn more about AI concepts and view examples in Examples and concepts.

Browse AI Workflow templates.

Read What product people need to know about LangChain to learn more about
AI concepts and terminology.

Was this page helpful?

Previous Next
Advanced AI Overview

Change cookie settings

Made with Material for MkDocs Insiders
n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Advanced AI LangChain in n8n

Tutorial: Build an AI workflow

in n8n LangChain in n8n
LangChain in n8n
Overview n8n provides a collection of nodes that implement LangChain's functionality. The
Langchain concepts in n8n LangChain nodes are configurable, meaning you can choose your preferred agent,
LangChain learning LLM, memory, and so on. Alongside the LangChain nodes, you can connect any n8n
resources node as normal: this means you can integrate your LangChain logic with other data
Examples and concepts sources and services.

Introduction
Learning resources: n8n's documentation for LangChain assumes you're familiar
What is a tool? with AI and LangChain concepts. This page provides links to learning resources.
Use Google Sheets as a data
source LangChain concepts and features in n8n: how n8n represents LangChain
Call an API to fetch data concepts and features.

Set a human fallback for AI

workflows Was this page helpful?

Previous Next
Tutorial: Build an AI workflow in n8n Langchain concepts in n8n

Change cookie settings

Made with Material for MkDocs Insiders
n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Advanced AI LangChain in n8n Table of contents

Tutorial: Build an AI workflow
in n8n LangChain concepts in n8n Trigger nodes
Cluster nodes
LangChain in n8n Root nodes
Overview This page explains how LangChain concepts and features map to n8n nodes. Chains
Langchain concepts in n8n Agents
This page includes lists of the LangChain-focused nodes in n8n. You can use any
LangChain learning Vector stores
resources n8n node in a workflow where you interact with LangChain, to link LangChain to
Miscellaneous
other services. The LangChain features uses n8n's Cluster nodes.
Examples and concepts Sub-nodes
Introduction n8n implements LangChain JS Document loaders
What is a tool? Language models
This feature is n8n's implementation of LangChain's JavaScript framework.
Use Google Sheets as a data Memory
source
Output parsers
Call an API to fetch data
Retrievers
Set a human fallback for AI Trigger nodes Text splitters
workflows
Chat Trigger Tools
Embeddings
Miscellaneous
Cluster nodes
Cluster nodes are node groups that work together to provide functionality in an n8n
workflow. Instead of using a single node, you use a root node and one or more sub-
nodes. This allows extensive configuration of node functionality.

Root nodes

Each cluster starts with one root node.

Chains

A chain is a series of LLMs, and related tools, linked together to support

functionality that can't be provided by a single LLM alone.

Available nodes:

Basic LLM Chain

Retrieval Q&A Chain

Summarization Chain

Learn more about Chains in LangChain.

Agents

An agent has access to a suite of tools, and determines which ones to use
depending on the user input. Agents can use multiple tools, and use the output of
one tool as the input to the next. Source

Available nodes:

Agent

Learn more about Agents in LangChain.

Vector stores

Vector stores store embedded data, and perform vector searches on it.

In Memory Vector Store

Pinecone Vector Store

Qdrant Vector Store

Supabase Vector Store

Zep Vector Store

Learn more about Vector stores in LangChain.

Miscellaneous

Utility nodes.

LangChain Code: import LangChain. This means if there is functionality you need
that n8n hasn't created a node for, you can still use it.

Sub-nodes

Each root node can have one or more sub-nodes attached to it.

Document loaders

Document loaders add data to your chain as documents. The data source can be a
file or web service.

Available nodes:

Default Document Loader

GitHub Document Loader

Learn more about Document loaders in LangChain.

Language models

LLMs (large language models) are programs that analyze datasets. They're the key
element of working with AI.

Available nodes:

Anthropic Chat Model

AWS Bedrock Chat Model

Cohere Model

Google PaLM Chat Model

Google PaLM Language Model

Hugging Face Inference Model

Mistral Cloud Chat Model

Ollama Chat Model

Ollama Model

OpenAI Chat Model

OpenAI Model

Learn more about Language models in LangChain.

Memory

Memory retains information about previous queries in a series of queries. For

example, when a user interacts with a chat model, it's useful if your application can
remember and call on the full conversation, not just the most recent query entered
by the user.

Available nodes:

Motorhead

Redis Chat Memory

Window Buffer Memory

Xata

Zep

Learn more about Memory in LangChain.

Output parsers

Output parsers take the text generated by an LLM and format it to match the
structure you require.

Available nodes:

Auto-fixing Output Parser

Item List Output Parser

Structured Output Parser

Learn more about Output parsers in LangChain.

Retrievers

Contextual Compression Retriever

MultiQuery Retriever

Vector Store Retriever

Workflow Retriever

Text splitters

Text splitters break down data (documents), making it easier for the LLM to process
the information and return accurate results.

Available nodes:

Character Text Splitter

Recursive Character Text Splitter

Token Splitter

n8n's text splitter nodes implements parts of LangChain's text_splitter API.

Tools

Utility tools.

Calculator

Code Tool

SerpAPI

Wikipedia

Wolfram|Alpha

Workflow Tool

Embeddings

Embeddings capture the "relatedness" of text, images, video, or other types of

information. (source)

Available nodes:

Embeddings AWS Bedrock

Embeddings Cohere

Embeddings Google PaLM

Embeddings Hugging Face Inference

Embeddings Mistral Cloud

Embeddings Ollama

Embeddings OpenAI

Learn more about Text embeddings in LangChain.

Miscellaneous

Chat Memory Manager

Was this page helpful?

Previous Next
Overview LangChain learning resources

Change cookie settings

Made with Material for MkDocs Insiders
n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Advanced AI LangChain in n8n

Tutorial: Build an AI workflow

in n8n LangChain learning resources
LangChain in n8n
Overview You don't need to know details about LangChain to use n8n, but it can be helpful to
Langchain concepts in n8n learn a few concepts. This pages lists some learning resources that people at n8n
LangChain learning have found helpful.
resources
The LangChain documentation includes introductions to key concepts and possible
Examples and concepts
use cases. Choose the LangChain | Python or LangChain | JavaScript
Introduction
documentation for quickstarts, code examples, and API documentation. LangChain
What is a tool? also provide code templates (Python only), offering ideas for potential use cases and
Use Google Sheets as a data common patterns.
source
Call an API to fetch data What Product People Need To Know About LangChain provides a list of terminology
Set a human fallback for AI and concepts, explained with helpful metaphors. Aimed at a wide audience.
workflows
If you prefer video, this YouTube series by Greg Kamradt works through the
LangChain documentation, providing code examples as it goes.

n8n offers space to discuss LangChain on the Discord. Join to share your projects
and discuss ideas with the community.

Was this page helpful?

Previous Next
Langchain concepts in n8n Introduction

Change cookie settings

Made with Material for MkDocs Insiders
n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Advanced AI Examples and concepts

Tutorial: Build an AI workflow

in n8n Advanced AI examples and concepts
LangChain in n8n
Overview This section provides explanations of important AI concepts, and workflow
Langchain concepts in n8n templates that highlight those concepts, with explanations and configuration guides.
LangChain learning The examples cover common use cases and highlight different features of advanced
resources AI in n8n.
Examples and concepts
Introduction Custom n8n Workflow Tool More concepts and examples
What is a tool? coming soon
Learn about tools in AI, then explore
Use Google Sheets as a data examples that use n8n workflows as n8n will add more examples in the
source
custom tools to give your AI workflow coming months. In the mean time,
Call an API to fetch data
access to more data. you can browse AI templates,
Set a human fallback for AI included community contributions,
workflows What is a tool in AI?
on the n8n website.
Chat with Google Sheets
Call an API to fetch data Browse all AI templates
Set up a human fallback

Was this page helpful?

Previous Next
LangChain learning resources What is a tool?

Change cookie settings

Made with Material for MkDocs Insiders
n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Advanced AI Examples and concepts Table of contents

Tutorial: Build an AI workflow
in n8n What is a tool in AI? Tools in n8n

LangChain in n8n
Overview In AI, 'tools' has a specific meaning. Tools act like addons that your AI can use to
Langchain concepts in n8n access extra context or resources.
LangChain learning
resources Here are a couple of other ways of expressing it:

Examples and concepts Tools are interfaces that an agent can use to interact with the world (source)
Introduction
We can think of these tools as being almost like functions that your AI model can
What is a tool?
call (source)
Use Google Sheets as a data
source
Call an API to fetch data
Set a human fallback for AI
Tools in n8n
workflows
n8n provides tool sub-nodes that you can connect to your AI agent. As well as
providing some popular tools, such as Wikipedia and SerpAPI, n8n provides two
especially powerful tools:

Custom n8n Workflow Tool: use this to load any n8n workflow as a tool.

Custom Code Tool: write code that your agent can run.

The next three examples highlight the Custom n8n Workflow Tool:

Chat with Google Sheets

Call an API to fetch data

Set up a human fallback

Was this page helpful?

Previous Next
Introduction Use Google Sheets as a data source

Change cookie settings

Made with Material for MkDocs Insiders
n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Advanced AI Examples and concepts Table of contents

Tutorial: Build an AI workflow
in n8n Chat with a Google Sheet using AI Key features
Using the example
LangChain in n8n
Overview Use n8n to bring your own data to AI. This workflow uses the Chat Trigger to provide
Langchain concepts in n8n the chat interface, and the Custom n8n Workflow Tool to call a second workflow
LangChain learning that queries Google Sheets.
resources

Examples and concepts

Introduction
What is a tool?
Use Google Sheets as a data
source
Call an API to fetch data
Set a human fallback for AI
workflows

Download the example workflow

Key features
This workflow uses:

Chat Trigger: start your workflow and respond to user chat interactions. The
node provides a customizable chat interface.

Agent: the key piece of the AI workflow. The Agent interacts with other
components of the workflow and makes decisions about what tools to use.

Custom n8n Workflow Tool: plug in n8n workflows as custom tools. In AI, a tool is
an interface the AI can use to interact with the world (in this case, the data
provided by your workflow). It allows the AI model to access information beyond
its built-in dataset.

Using the example

To load the template into your n8n instance:

1. Download the workflow JSON file.

2. Open a new workflow in your n8n instance.

3. Copy in the JSON, or select Workflow menu > Import from file....

The example workflows use Sticky Notes to guide you:

Yellow: notes and information.

Green: instructions to run the workflow.

Orange: you need to change something to make the workflow work.

Blue: draws attention to a key feature of the example.

Was this page helpful?

Previous Next
What is a tool? Call an API to fetch data

Change cookie settings

Made with Material for MkDocs Insiders
n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Advanced AI Examples and concepts Table of contents

Tutorial: Build an AI workflow
in n8n Call an API to fetch data Key features
Using the example
LangChain in n8n
Overview Use n8n to bring data from any API to your AI. This workflow uses the Chat Trigger to
Langchain concepts in n8n provide the chat interface, and the Custom n8n Workflow Tool to call a second
LangChain learning workflow that calls the API. The second workflow uses AI functionality to refine the
resources API request based on the user's query.
Examples and concepts
Introduction
What is a tool?
Use Google Sheets as a data
source
Call an API to fetch data
Set a human fallback for AI
workflows

Download the example workflow

Key features
This workflow uses:

Chat Trigger: start your workflow and respond to user chat interactions. The
node provides a customizable chat interface.

Agent: the key piece of the AI workflow. The Agent interacts with other
components of the workflow and makes decisions about what tools to use.

Custom n8n Workflow Tool: plug in n8n workflows as custom tools. In AI, a tool is
an interface the AI can use to interact with the world (in this case, the data
provided by your workflow). It allows the AI model to access information beyond
its built-in dataset.

A Basic LLM Chain with an Auto-fixing Output Parser and Structured Output
Parser to read the user's query and set parameters for the API call based on the
user input.

Using the example

To load the template into your n8n instance:

1. Download the workflow JSON file.

2. Open a new workflow in your n8n instance.

3. Copy in the JSON, or select Workflow menu > Import from file....

The example workflows use Sticky Notes to guide you:

Yellow: notes and information.

Green: instructions to run the workflow.

Orange: you need to change something to make the workflow work.

Blue: draws attention to a key feature of the example.

Was this page helpful?

Previous Next
Use Google Sheets as a data source Set a human fallback for AI workflows

Change cookie settings

Made with Material for MkDocs Insiders
n8n home Forum Blog

n8n Docs Search

Using n8n Integrations Hosting n8n Code in n8n Advanced AI API Embed

Advanced AI Advanced AI Examples and concepts Table of contents

Tutorial: Build an AI workflow
in n8n Have a human fallback for AI workflows Key features
Using the example
LangChain in n8n
Overview This is a workflow that tries to answer user queries using the standard GPT-4 model.
Langchain concepts in n8n If it can't answer, it sends a message to Slack to ask for human help. It prompts the
LangChain learning user to supply an email address.
resources
This workflow uses the Chat Trigger to provide the chat interface, and the Custom
Examples and concepts
n8n Workflow Tool to call a second workflow that handles checking for email
Introduction
addresses and sending the Slack message.
What is a tool?
Use Google Sheets as a data
source
Call an API to fetch data
Set a human fallback for AI
workflows

Download the example workflow

Key features
This workflow uses:

Chat Trigger: start your workflow and respond to user chat interactions. The
node provides a customizable chat interface.

Agent: the key piece of the AI workflow. The Agent interacts with other
components of the workflow and makes decisions about what tools to use.

Custom n8n Workflow Tool: plug in n8n workflows as custom tools. In AI, a tool is
an interface the AI can use to interact with the world (in this case, the data
provided by your workflow). It allows the AI model to access information beyond
its built-in dataset.

Using the example

To load the template into your n8n instance:

1. Download the workflow JSON file.

2. Open a new workflow in your n8n instance.

3. Copy in the JSON, or select Workflow menu > Import from file....

The example workflows use Sticky Notes to guide you:

Yellow: notes and information.

Green: instructions to run the workflow.

Orange: you need to change something to make the workflow work.

Blue: draws attention to a key feature of the example.

Was this page helpful?

Previous Next
Call an API to fetch data n8n public REST API

Change cookie settings

Made with Material for MkDocs Insiders