n8n AI Stack: How to Deploy a Full AI Automation System in 15 Minutes
I’ve been messing around with automation and AI for years. Everything from simple Telegram bots to huge RAG setups with all the bells and whistles. And there’s one thing I’ve always hated:
Setting up new projects. Every. Single. Time.
It felt like I was throwing hours into a black hole. Sometimes, I’d lose a whole day just trying to get the environment ready to go.
First, it was Docker Compose for n8n. Then, PostgreSQL. Oh, and don’t forget Redis – because n8n just won’t work in queue more without it. After that, Supabase for vectors… then Qdrant, because Supabase can be slow when dealing with them. And HTTPS, of course! Always HTTPS. Caddy or Nginx, getting certificates, fixing issues… Over and over.
I remember one setup where I thought, This is crazy; there has to be a better way.
That’s how n8n-install was born – a repo that turns a fresh Ubuntu server into a fully operating AI stack with a single command.
Once you've deployed your n8n AI stack, you might want to consider how to efficiently host it, and we have a guide for that, showing you how to Run n8n and Supabase on One VPS: Ultimate Coolify Setup Guide.
Table of Contents
What Is an n8n AI Stack
n8n-install isn’t just another Docker Compose file. It’s a full setup ready to make things happen. It’s a ready-made way to get a complete automation and AI setup going.
At the center is n8n – a low-code tool with a ton of integrations and great AI support. But it’s not just about n8n, it’s a bunch of things working together.
SPONSORED
Get back 10–15 hours a week using AI—no coding skills or crazy promises needed.
Join our 3-hour AI Workflow Makeover session where business owners create actual automations they can start using in two days.
You’ll build at least 3 workflows and get templates, plus lifetime access to the program.
You get a production-ready setup right away: HTTPS is set up, all necessary services are connected, and the network is ready. No need to mess with settings or worry about SSL. It just works.
Repository: GitHub
And hey, if you find it helpful, give it a ⭐. It really helps new open-source projects grow!
Why Deploy an n8n AI Stack
One of my recent projects needed a pretty complicated AI setup.
- n8n with Redis queues
- Flowise for managing LLMs
- Qdrant for vector storage
- PostgreSQL for keeping data safe
- Ollama for using models locally
- Crawl4AI for grabbing data
- Grafana for keeping an eye on things
- Postgresus for backups
- LangFuse for understanding AI performance
- Portainer for managing containers
Setting all of this up on a new server was exhausting. I was wasting time, looking at old notes and copying files, changing ports and messing with settings. Something always went wrong – a typo, a missing setting, or some SSL issue.
I realized I was wasting my time on stuff a script could do.
So I created an installer that takes care of everything.
What’s Included in the n8n AI Stack
You choose what you want to install with a simple setup, and there are over 30 options.
Main Stuff
- PostgreSQL + Redis – data storage and managing tasks
- Caddy – HTTPS with SSL certificates
AI Tools
- n8n – runs workflows in parallel
- Ollama – run local models (like Llama and Mistral)
- Open WebUI – a ChatGPT-like interface for own models
- Flowise – an AI builder that works well with n8n
- Dify – a platform for building AI applications
RAG and Vector Storage
- Supabase – Postgres, vectors, and user management all in one.
- Qdrant – fast vector storage
- Weaviate – another vector option
- RAGApp, RAGFlow, LightRAG – RAG setups
Monitoring and Control
- LangFuse – see how your AI and LLM are doing
- Grafana + Prometheus – keep an eye on your setup
- Portainer – manage Docker through a web interface
- Postgresus – Postgres monitoring and backups
Extra Tools
- ComfyUI – Stable Diffusion workflows
- Neo4j – graph database
- Crawl4AI – scrape websites
- PaddleOCR – get text from images
- LibreTranslate – self-host translate
- Python Runner – run Python code from n8n
You don’t need everything! Just get what you need. n8n, Ollama, and Qdrant is a great start
Required Tools for n8n AI Stack Setup
It’s pretty simple.
You’ll need:
- A domain (a cheap one is fine)
- A server running Ubuntu
Create a DNS record (*.yourdomain.com) that points to your server. Then, connect via SSH and run:
git clone https://github.com/kossakovsky/n8n-installcd n8n-installsudo bash ./scripts/install.shThe installer will ask for:
- Your domain
- Email (for SSL and login details)
- OpenAI API key (if you want it)
- Whether to import n8n workflows
- Number of n8n workers
- What services to install
That’s it.
In about 10–15 minutes, you’ll have a secured setup. It will take longer if you load workflows.
n8n.yourdomain.com– up and runningflowise.yourdomain.com– up and runningwebui.yourdomain.com– up and running
All these things work together on a Docker network. n8n can use Ollama, store stuff in Qdrant, and read from Supabase. Flowise can start n8n workflows. It all works together.
Basic needs: Ubuntu 24.04, 4 GB RAM, 2 CPU cores, 30 GB disk. Remember to scale up if you install lots of things!
Extras I Added
Queue Mode with Workers
n8n usually runs things one at a time. But with queue mode, it uses Redis and multiple workers to do things at the same time.
Workflows Included
There are a lot of n8n workflows ready to use. RAG setups, AI agents, Gmail/Notion/Airtable integrations, document processing, bots for Telegram, Discord, Slack, and more. Start with these instead of starting from scratch.
More Code Options
JavaScript nodes have cheerio, axios, moment, and lodash. ffmpeg is also there for audio and video without a lot of setup.
HTTPS with Certificates
Caddy handles everything. Let’s Encrypt gives you certificates. You don’t have to worry about them!
Cloudflare Tunnel
Hide your server IP. No need to open ports 80 and 443. Get DDoS protection.
Example of n8n AI Workflow Automation
I created a Telegram bot that answers questions using data from a website.
How it works:
- n8n manages the bot.
- Flowise is the brains.
- Qdrant stores data.
- Ollama makes responses.
- Crawl4AI grabs data.
- PostgreSQL stores conversations and logs.
The path is User -> n8n -> Flowise -> Ollama -> n8n -> Telegram.
LangFuse monitors how it’s doing, Grafana tracks the system, Postgresus backs up, and Portainer shows everything.
Before, this took hours to set up. Now, it takes minutes for installation and a little bit extra for adding settings. The installer creates a report with addresses, credentials, and API keys.
SPONSORED
For deploying n8n-install, you need a fast and stable VPS.
Senko Digital offers AMD EPYC CPUs (up to 5.0 GHz), NVMe RAID10 storage, up to 1 Gbps network, root access, and activation in under 2 minutes.
2–4 vCPU is enough for a basic stack.
For AI + RAG workloads, go with 4–8 vCPU.
Who It’s For
If you:
- Make AI agents or chatbots
- Use n8n (or want to)
- Want a private AI setup
- Don’t want to waste time on Docker and SSL
- Want fast setups
Then this is for you.
Hope I saved your time.
Source Author: Yury Kossakovsky
