Lesson 4 — Building Your First Local AI Chatbot

Maracade Technologies · AI Agent Course

Lesson 4 — Building Your First
Local AI Chatbot

Learn to build a fully local AI assistant using n8n, Ollama, and Mistral — running entirely on your own machine.

⚙️ n8n Workflows

🦙 Ollama

🤖 Mistral AI

💬 Chat Trigger

🔒 100% Local

▶ Start Lab (14 Steps)

Hands-on Visual Lab

Follow Each Step

Step 1 of 14

Search for Chat Trigger

Create a new workflow in n8n and use the trigger search panel to find the Chat node. This is where your chatbot will receive incoming messages.

localhost:5678/workflow/new · n8n Editor

＋

Add first step...

What triggers this workflow?

🔍 chat

💬 Chat

🤖 AI Agent

⚙️ OpenAI

📊 Chat Data

🧠 Chat Memory Manager

💡In the n8n editor, click "Add first step" then type "chat" in the trigger search box. The Chat trigger will appear at the top of the results list.

Step 2 of 14

Open Chat Trigger

Click the Chat item from the search results to open the trigger settings panel. You'll see the trigger options and actions available.

n8n · Chat Trigger Settings

💬 Chat

🔍 Search Chat Actions...

⚡ Triggers (1)

💬 On new Chat event ⚡

Actions (2)

💡Select "On new Chat event" under Triggers. This will fire your workflow every time someone sends a message to your chatbot.

Step 3 of 14

Configure Chat Trigger

Open the Chat trigger settings tab. You'll see options like Make Chat Publicly Available and Make Available in n8n Chat Hub. Leave defaults for now and proceed.

n8n · When chat message received · Parameters

Parameters Settings

🔗 Open chat

Make Chat Publicly Available

Make Available in n8n Chat Hub

Options

No properties

⚡

No trigger output

Test this trigger

or set mock data

Step 4 of 14

Add Message a Model Node

Connect the Message a Model node to the Chat Trigger. This node handles the communication with the AI model. You'll see it appear in the workflow canvas connected to the trigger.

n8n · Workflow Canvas

💬

When chat message

received

──▶

🦙

Message a

model

message: text

Tools

──▶ ＋

💡Click the + connector dot on the right side of the Chat Trigger node, then search for "Message a Model" to add it to the canvas.

Step 5 of 14

Open Message a Model

Double-click the Message a Model node to open its configuration screen. You'll need to set credentials, choose a model, and configure messages.

n8n · Message a model · Configuration

Parameters Settings

⚡ Execute step

Credential

No credentials yet

Set up credential

Resource

Text ▾

Operation

Message a Model ▾

Model

From list ▾

Choose... ▾

⚠️

Messages — Content

e.g. Hello, how can you help me?

Step 6 of 14

Create Ollama Credentials

Click "Set up credential" and configure the Ollama connection using the base URL http://localhost:11434 — no API key is required for local Ollama installations.

n8n · Ollama Account Setup

🦙

Ollama account

Ollama

Save

Connection

Sharing

Details

Need help? Read our docs

Base URL *

http://localhost:11434

API Key

(not required for local)

🔑Ollama runs locally, so no API key is needed. The default port is 11434. Make sure Ollama is running on your machine before saving.

Step 7 of 14

Save Ollama Account

Click Save to store the Ollama credentials. n8n will test the connection and confirm it's working. Your Ollama account will appear in the credentials dropdown.

n8n · Credentials Saved

🦙

Ollama Account Connected

✅ Connection verified

Credential Name

Ollama account

Base URL

http://localhost:11434

Step 8 of 14

Load AI Model

With Ollama credentials saved, click the Model dropdown. n8n will query your local Ollama installation and list all available models. Select mistral from the list.

n8n · Select AI Model

Model

From list ▾

mistral ▾

✅ mistral

llama3.2

codellama

phi3

💡If you don't see mistral in the list, you need to pull it first. Open a terminal and run: ollama pull mistral

Step 9 of 14

Run Mistral in PowerShell

Test that Mistral is working correctly by running it directly from the command line using Ollama. This confirms the model is installed and responding before connecting it to n8n.

Windows PowerShell

Install the latest PowerShell for new features: https://aka.ms/PSWindows

C:\Users\HP> ollama run mistral

>>> hello

Hello! How can I help you today? Is there something specific you'd like to talk about or learn more about? I'm here to assist you with information, answer questions, or just have a conversation. Let me know!

⚡Type ollama run mistral in PowerShell to start an interactive session. Type /bye to exit when done testing.

Step 10 of 14

Add Chat Response

Connect a Chat response node to the output of the Message a Model node. This node sends the AI's response back to the chat window so users see the reply.

n8n · Add Response Node

What happens next?

🔍 respond

🔗 Respond to Webhook Returns data for Webhook

✅ Respond.io Read, update, write...

💬 Chat →

More from the community

Step 11 of 14

Configure Chat Message

Open the Chat response node and configure it. Set the Operation to Send Message and map the Message field to the AI model's output. Make sure Response Mode is set to Using Response Nodes.

n8n · Chat · Configuration

⚠️ Verify you're using a chat trigger with the 'Response Mode' option set to 'Using Response Nodes'

Operation

Send Message ▾

Message

{{ $json.content }}

Options

No properties

🔗Map the Message field to {{ $json.content }} to pass the AI's response back to the chat interface.

Step 12 of 14

Open Chat Window

Click the orange "Open chat" button in n8n to launch the built-in chat interface at the bottom of the editor. This is where you'll test your chatbot.

n8n · Chat Interface Open

💬

When chat message

received

──▶

🦙

Message a model

message: text

──▶ ＋

🔗 Open chat

💬 Chat Session: 0322a... 📋 Logs

Step 13 of 14

Execute Workflow

Type a test message in the chat window and press Enter. Watch the workflow execute in real-time — you'll see green checkmarks appear on each node as the message flows through and Mistral processes your request.

n8n · Workflow Executing...

💬

1 item ✅

When chat message

received

──▶

🦙

Message a model

⏳ Processing...

Type message, or press 'up' for...

Step 14 of 14

✅ Success Result

Your chatbot successfully receives the message, sends it to Mistral via Ollama, and returns the AI-generated response. The execution log shows Success in 7.559s and the chat displays the AI's reply.

n8n · Workflow Execution Complete ✅

💬

1 item ✅

When chat message

received

──▶

🦙

Message a model

1 item ✅

──▶ ＋

✅ Success in 7.559s

{ "content": " Hello! How can I assist you today?" }

🎉Congratulations! Your local AI chatbot is working. The response came entirely from your machine using Ollama + Mistral — no internet, no API costs, total privacy.

🏆

Lesson 4 Complete!

You've built a fully local AI chatbot from scratch. Here's what you mastered:

```

⚙️

n8n Workflows
Visual automation pipelines

🦙

Ollama Integration
Local AI model management

🤖

Mistral AI Model
Open-source LLM running locally

💬

Chat Systems
Trigger → Process → Respond

🔒

Privacy-First AI
No data leaves your machine

Coming Up Next

🧠 AI Memory Systems

📱 WhatsApp AI Bots

🌐 Website AI Chatbots

📊 CRM Automation

🎧 AI Customer Support

💼 AI Agent Rental Business

```

Lesson 4 — Building Your FirstLocal AI Chatbot

Lesson 4 — Building Your First
Local AI Chatbot