DEV Community

Cover image for Building a Self-Hosted n8n AI Automation Lab with Ollama & PostgreSQL on Windows 🤖💻
Izuchukwu Nwachukwu
Izuchukwu Nwachukwu

Posted on

Building a Self-Hosted n8n AI Automation Lab with Ollama & PostgreSQL on Windows 🤖💻

#n8n #docker #automation #selfhosted

As a tech enthusiast, I’ve always been fascinated by the potential of AI automation. This week, I decided to roll up my sleeves and build something powerful; a fully self-hosted AI automation lab, right from my Windows HP PC.
Using the n8n Self-hosted AI Starter Kit, I configured a local stack that combines:

🧠 Ollama, which powers the local AI engine by running large language models like LLaMA 3 directly on my machine — fast, private, and cloud-free

🗃️ PostgreSQL, serving as n8n’s database to reliably store workflows, credentials, and execution logs

⚙️ n8n as the automation engine and workflow editor

🧰 Docker, Git Bash, and VS Code to orchestrate and customize the entire environment

📊 Qdrant → vector database for semantic search and memory

Here's why self-hosting n8n caught my attention:

  • No Subscription Fees: Enjoy the features of n8n without recurring costs.
  • Total Control: Manage your data, workflows, and security with ease.
  • Scalability: Run unlimited workflows without restrictions.
  • Customization: Tailor n8n to fit your specific needs with custom nodes and integrations.

In this article, I'll share my experience setting up my self-hosted n8n AI automation lab and walk you through the process.

✅ Prerequisites

Make sure you have:

  • Docker and Docker Compose v2+
  • Git
  • Terminal access
  • Web browser
  • At least 4GB RAM (8GB+ recommended) and 10GB+ disk space

My Setup and Tools

  • OS: Windows 10/11 PC
  • Terminal: Git Bash
  • Editor: Visual Studio Code
  • Docker: via Docker Desktop
  • Browser: Chrome (for n8n dashboard)
  • Hardware: Intel HD Graphics 520 (CPU-only)

Configuration Guide

Step 1 – Clone the repo 

git clone https://github.com/n8n-io/self-hosted-ai-starter-kit.git
cd self-hosted-ai-starter-kit
Enter fullscreen mode Exit fullscreen mode

First-time setup showing Docker pulling required images and creating containers

First run: Docker pulls required images, creates volumes/network and creates containers.

Database migrations and startup logs while initializing n8n

Database migrations and n8n-import logs showing tables and migrations being applied

Step 2 – Configure .env

I opened .env in VS Code and added my secrets (dummy values were used for sensitive keys):

env
POSTGRES_USER="root"
POSTGRES_PASSWORD="examplepass123"
POSTGRES_DB="n8n"

N8N_ENCRYPTION_KEY="examplekey123"
N8N_USER_MANAGEMENT_JWT_SECRET="examplejwtsecret123"
N8N_DEFAULT_BINARY_DATA_MODE="filesystem"

N8N_RUNNERS_ENABLED=true
N8N_BLOCK_ENV_ACCESS_IN_NODE=false
N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS=true

💡 Pro tip: Always quote your secrets and escape $ signs, or Docker will misinterpret them.

Step 3 – Start the Setup 

docker compose --profile cpu up
Enter fullscreen mode Exit fullscreen mode

Re-running docker compose --profile cpu up, confirming services (Postgres, Qdrant, Ollama, n8n) are re-initialized

Re-running docker compose --profile cpu up, confirming services (Postgres, Qdrant, Ollama, n8n) are re-initialized

This spun up:

  • n8n
  • PostgreSQL
  • Qdrant
  • Ollama
  • n8n-import

⚠️ Common Errors and Fixes

  • ** Environment Variable Warnings** → fixed by quoting + escaping $
  • ** Network conflicts** → stopped all containers and removed the old network
  • n8n stuck on startup → checked logs, restarted, and it finally recovered

✅ Accessing the Dashboard

Once everything was up, I visited:

http://localhost:5678
Enter fullscreen mode Exit fullscreen mode

…and boom 💥 — the n8n editor was live!

n8n dashboard running on 5678

🔄 Restarting after Reboots

Each time I turn on my PC, I simply run:

cd ~/self-hosted-ai-starter-kit
docker compose --profile cpu up
Enter fullscreen mode Exit fullscreen mode

🔖 Optional Startup Script
To make things easier, I created a file called start-n8n.sh:

#!/bin/bash
cd ~/self-hosted-ai-starter-kit
docker compose --profile cpu up
Enter fullscreen mode Exit fullscreen mode

Then I ran:

./start-n8n.sh
Enter fullscreen mode Exit fullscreen mode

You may need to make it executable:

chmod +x start-n8n.sh
Enter fullscreen mode Exit fullscreen mode

🌍 For Others Testing This

  • Mac (Apple Silicon) 
  docker compose --profile mac up
Enter fullscreen mode Exit fullscreen mode
  • Linux (AMD GPU) 
  docker compose --profile gpu-amd up
Enter fullscreen mode Exit fullscreen mode
  • Nvidia GPU 
  docker compose --profile gpu-nvidia up
Enter fullscreen mode Exit fullscreen mode

What’s Next

Setting up my self-hosted AI automation lab with n8n, Ollama, and PostgreSQL has been a rewarding experience. It gave me a reliable and flexible platform to build intelligent workflows without relying on cloud services or subscriptions. Everything runs locally, giving me full control over my data and processes.

Whether you're building smart assistants, automating tasks, or experimenting with local LLMs, this starter kit is a great place to begin. If you're curious about how it works or want to dive deeper, I recommend checking out n8n’s official documentation. It’s clear, practical, and built to support creators at every level.

I’ll be sharing more soon as I build my next n8n project. Feel free to follow, engage, and stay tuned for what’s coming next.

Top comments (0)