Self-Hosting n8n: How to Save Hundreds on Automation (and Get More Power)
A while back, a friend was paying $29/month for Make.com.
It worked — but once he started automating more, the limits hit fast. 10,000 ops went quicker than expected. Then came the upgrade prompts, the usage warnings, and eventually, the $50/month plan.
That’s when he asked a very simple question:
Could I just run this myself?
Turns out, yeah - he could. And now he runs n8n on a $5/month server, with no limits, no surprise charges, and full control. Better yet? He also runs his own vector DB, video transcription, and API glue — all on the same box.
If you're building AI agents, automating your business, or just tired of SaaS bloat — here's how self-hosting n8n (and friends) actually works, and why it might be the best technical decision you make this year.
Why Self-Hosting Is Way More Accessible Than You Think
Let’s kill the myth right away:
You don’t need to be a backend dev to self-host n8n.
All you really need is:
- A VPS (a virtual private server — like a tiny rented computer in the cloud)
- Docker (a way to run apps without installing all their stuff manually)
You spin up the VPS (e.g. Hetzner, DigitalOcean or Racknerd), install Docker, and run the official n8n container. That’s it.
From there, you’ve got your own automation platform. No rate limits. No “premium” steps. Just a blank canvas that runs as much and as often as you want.
So… Why Bother?
Two reasons:
- It’s way cheaper.
- You get more control.
Here’s a quick cost breakdown:
- Make.com → $9/month → ~10k ops
- n8n Cloud → $20/month → ~10k workflow executions (unlimited ops)
- Self-hosted n8n → $5/month → unlimited executions
Even if you spin up a backup VPS for $20/year, you’re still well under $10/month for something that could run your entire automation backend.
But here’s the catch most people miss:
As soon as you go from "a few zaps" to serious automations — scheduling agents, processing large data, handling thousands of triggers daily — both Make and n8n Cloud can easily run into the hundreds per month.
That’s where self-hosting flips from “clever hack” to necessary infrastructure.
And it’s not just about n8n…
You’re Building a Whole Stack — Not Just Hosting One Tool
Here’s where it gets fun.
Once you’ve got that $5/month VPS, it becomes a platform, not just a container for n8n.
On mine, I run:
- Whisper to transcribe videos locally
- Qdrant as a private vector database
- PostgreSQL for persistent data — no Supabase subscription needed
- A small FastAPI app that exposes tools like Whisper via HTTP endpoints
Then I call those endpoints directly from n8n. All my tools — connected through one clean automation layer, running on my own terms.
Bonus: no usage-based billing. Everything I add is just more value squeezed out of the same $5 server.
The Tradeoffs (Yes, There Are a Few)
Self-hosting gives you power, but it also gives you responsibility. Here’s what you’ll want to handle:
Basic Security
Your n8n instance is exposed to the internet — secure it:
- Use strong authentication
- Set up a firewall (UFW or provider-based)
- Add HTTPS (Let’s Encrypt + reverse proxy like Traefik)
Backups
Your workflows live on that machine. If it dies, they’re gone.
I back up my n8n data to a separate VPS nightly (costs ~$20/year). You can script this with GPT in minutes.
Updates
Every now and then, you’ll pull the latest Docker images and restart. Takes two commands. Or automate it with something like watchtower.
Nothing here is hard — but it's different from clicking “Upgrade Plan.”
When It’s Worth It — And When It’s Not
Self-hosting isn’t for everyone.
But here’s when it definitely makes sense:
- You’re building agents that run 24/7
- You want to avoid usage-based billing traps
- You’re connecting multiple tools with complex logic
- You’re tired of rebuilding workflows because a SaaS tool changed its limits
It might not be worth it if you’re just testing ideas, or if you really don’t want to think about infrastructure. But if your automations are part of how your business runs? It’s worth the jump.
Final Thought: This Is What Building Leverage Looks Like
I started self-hosting to save money.
But what I ended up with was leverage: a stack I control, tools I can extend, and automations I can run at any scale — all without asking permission or checking my usage stats.
If you’re using GPT or Make or n8n Cloud and feel like you’re starting to outgrow the limits — trust that instinct. You don’t need to downgrade your workflows. You just need to own the infrastructure behind them.
Want help setting up your own stack — or fixing the one GPT half-built for you?
Let’s talk.