N8N Part 1

Introduction

We're going to set up the AI Agent. I will be on my server, but this can easily be done on your local machine.

Problem

We need to create AI agents to help automate mundane tasks.
Need: Prosgresql, npm, N8N, Ollama

Plan

Install dependencies and get them all connected.
Get a simple Agent working
Connect with Cloudflared

Procedure

Let's get right into it! It appears I already have an Ubuntu machine with Docker installed. Let's update/upgrade and use this I noticed I only have 32GB of storage on this machine, and the RAM and CPU power were all low, so I upped it before I do the update and upgrade I doubled the RAM, CPU power, and upped the storage to 60GB. I also have to partition the extra storage. To adjust the partition:

//Check the current size:
df -h
sudo apt install cloud-guest-utils -y
sudo growpart /dev/sda 3 
// double chekc hte size is now 60GB
df -h

Got that taken care of, let's install PostgreSQL

docker run --name pgvector-container -e POSTGRES_USER=myuser -e POSTGRES_PASSWORD=mypassword -e POSTGRES_DB=mydatabase -p 5432:5432 -d ankane/pgvector

Install Ollama (Linux) (I had to install curl first)

curl -fsSL <https://ollama.com/isntall.sh> | sh

Download the model (it’s different for each model)

ollama run llama3.2

install Node.js

sudo apt update
sudo apt install -y curl software-properties-common
curl -fsSL <https://deb.nodesource.com/setup_20.x> | sudo -E bash -
sudo apt install -y nodejs

Now I don't have a GPU in my PowerEdg,e so ill be running this in CPU mode only. Lets see how taxing it will be on 8 cores. Once everything is installed/downloaded, let's start the N8N process

npx n8n start

It is really that easy, at the bottom, it will tell you the editor is accessible at

<http://localhost:5678>

I can just add this to my list of CloudFlare tunnels and edit from anywhere. I set it as HTTP://192.168.1.19:5678 == n8n.ashtonward.tech and then logged in and was connected.