The Real Cost of Self-Hosting n8n on a VPS (vs. One Flat Price)

Self-hosting n8n is genuinely appealing. You own your data, there's no vendor lock-in, and the software itself is free. But "free software on a cheap VPS" is not the same as a free total cost. By the time you've configured Docker, reverse-proxied Nginx, sorted out SSL, wired in a metered AI API, and set up nightly backups, you've spent money and hours that rarely make it into a self-hosting calculator.

This post adds all of it up, honestly, so you can decide with your eyes open.

---

The Line Items People Usually Leave Out

1. The VPS

A basic Hetzner or DigitalOcean VPS that can run n8n without choking on concurrent executions is roughly $6–12/month. The very cheapest shared VPS works until you have more than a couple of active workflows — the main process is single-threaded by default, and queue mode needs a Redis sidecar on the same box or an extra one.

Realistic starting point for anything beyond toy workflows: $8–10/month (2 vCPU, 4 GB RAM).

2. Docker, Nginx, and SSL

The official self-host guide uses Docker Compose. That means:

• Installing Docker and Docker Compose on the VPS
• Writing a docker-compose.yml for n8n + Postgres (SQLite is fine for small use, but Postgres is recommended for anything you care about)
• Adding an Nginx reverse proxy with Let's Encrypt certs via Certbot or Caddy
• Pointing a domain or subdomain at the server

None of this is rocket science, but it takes 2–4 hours the first time if you know what you're doing, and longer if you're learning on the job. Set a conservative hourly value for your time — even $30/hr — and that's $60–$120 of your hours just on initial setup.

3. Ongoing Maintenance

Self-hosted software doesn't update itself. n8n releases a new minor version roughly every two weeks. On a self-managed VPS you need to:

• Pull the new Docker image (docker compose pull && docker compose up -d)
• Watch for breaking changes in the release notes (some minor versions do break things — n8n has a long breaking-changes list)
• Rotate Let's Encrypt certs (Certbot automates this, but you need to verify it's actually working)
• Monitor disk space — workflow execution logs grow quietly
• Handle VPS reboots and ensure Docker containers restart correctly (restart: unless-stopped in Compose, but OS-level issues still require attention)

Realistically, 30–60 minutes per month at minimum to stay current and not wake up to a broken instance. Over 12 months: 6–12 hours/year of maintenance time.

4. The AI/LLM Bill — the Hidden Line Item

This is the one that catches most self-hosters by surprise. n8n's AI nodes (the OpenAI node, the LangChain agent, the embeddings node) all call external LLM APIs. You pay for those separately, on top of your VPS.

OpenAI GPT-4o runs at roughly $5 per million input tokens. A typical AI workflow that summarises emails, classifies tickets, or drafts content might consume 50,000–200,000 tokens per day depending on volume. That's:

• Light use (50k tokens/day): ~$7–10/month for the API alone
• Moderate use (200k tokens/day): ~$28–40/month for the API alone

This is in addition to the VPS. Total for a real self-hosted setup with AI nodes actively used: $35–55/month in hard costs, plus your ongoing time.

5. Backups

Postgres backups don't happen automatically. You need a cron job or a tool like pg_dump piped to S3 or Backblaze. Setting this up correctly (testing restores, not just dumps) is another 1–2 hours of work and potentially a few dollars a month in object storage.

Losing your workflows because of a disk failure on a VPS is not theoretical — it happens. Without a tested backup, your months of workflow-building go with it.

---

Putting Numbers Together

Cost item	DIY VPS	Notes
VPS rent	$8–10/mo	2 vCPU / 4 GB, Hetzner or DO
Domain/subdomain	~$1/mo	Amortised annual cost
LLM API (moderate use)	$15–40/mo	OpenAI / Anthropic, pay-per-token
Your setup time	$60–120 one-off	At $30/hr, conservative
Your maintenance time	$15–30/mo	30–60 min/mo at $30/hr
Backups	$1–3/mo	S3 / Backblaze
Total (first month)	$100–204	Including setup time
Total (steady state)	$40–84/mo	Excluding setup, including maintenance hours

These are conservative estimates. If you value your time more, or if you hit a tricky upgrade, the numbers go up.

---

What You Trade Off with DIY

To be fair: self-hosting gives you things that a hosted service doesn't.

• Full root access — you can install custom npm packages, mount volumes, run additional containers alongside n8n.
• No data ever leaves your server — your workflow data, credentials, and execution logs stay on hardware you control.
• No subscription dependency — you can freeze an n8n version indefinitely.

These are real advantages. If you have strong compliance requirements, a DevOps background, or need deep customisation, self-hosting is the right call and the costs above are worth paying.

The honest question is: do you need that level of control, or do you just want n8n to work?

---

The Alternative: One Flat Price, No Surprises

AgentRoost gives you your own n8n instance — your login, your workflows, your data — on a public subdomain (https://<your-id>.agentroost.app). You build workflows the same way. You export them the same way. The n8n editor is identical.

What's different:

• No Docker, no Nginx, no SSL config. The instance is live in about 2 minutes after signup.
• AI/LLM credits are included in the subscription. The AI nodes (LangChain agent, OpenAI, embeddings) already have credits wired in. No API key from OpenAI, no separate bill. This is uncommon — most n8n hosting options are bring-your-own-key. On AgentRoost, the AI works out of the box.
• 350+ LLM models accessible from the same included credit pool. Switch models anytime from your workflow without touching settings.
• Updates and backups are handled. No maintenance window you have to own.
• Flat price from $19.99/month — server, setup, and AI credits in one line item. Monthly billing, cancel anytime, 14-day money-back guarantee.

You trade root-level control for predictability and zero DevOps. Whether that trade is worth it depends on what you're building and how you value your time.

---

How to Get Your Own n8n Instance on AgentRoost

1. Sign up with email/password, Google, Microsoft, or Discord.
2. Pick the n8n framework from the agent selection screen.
3. Name your instance — this becomes your subdomain.
4. Your private n8n editor opens at https://<your-id>.agentroost.app. Log in with the credentials shown.
5. Build a workflow. Drop in an AI node. It already has credits — no API key prompt.

Webhooks get a real public HTTPS URL immediately. Credentials you save stay in your instance only.

If you want to evaluate before committing: the pricing page has a plan comparison, and there's a 14-day money-back guarantee if it turns out not to be what you need.

---

Which Should You Pick?

Self-host on a VPS if: you have DevOps experience, need root access for custom packages or adjacent containers, have hard data-residency requirements, or want to freeze a specific n8n version indefinitely.

Use AgentRoost if: you want n8n to work without becoming a system administrator, you want AI nodes that run without a separate OpenAI account, and you'd rather spend the setup hours on building workflows instead.

The numbers above are the honest version of what both choices cost. Make the call that fits your situation.

Compare plans →