n8n self-host vs Cloud cost calculator

Plug in your monthly executions, active workflows, and a rough hourly rate for the ops time self-hosting will eat. The calculator picks the cheapest n8n Cloud tier that fits, prices a self-hosted VPS stack at your volume, and tells you which one wins at your numbers - and at what volume the answer flips.

One full run of one workflow = 1 execution.
Workflows currently turned on. Drives Cloud tier limits.
Affects RAM sizing on self-host (memory per worker).
Used to value the ops time self-hosting takes (1-3 hrs/mo).
Auto: on at >50k executions/mo or workflows that hit concurrency.

Pricing verified 2026-05-08. Sources: n8n.io/pricing, Hetzner Cloud, DigitalOcean Droplets. Re-verify before quoting; pricing changes.

How the numbers are calculated

n8n Cloud tier selection

The calculator picks the cheapest published self-serve plan that fits both your execution count and your active-workflow count. Above the Pro cap, it shows "Enterprise (custom pricing)" rather than a fabricated number - n8n quotes those individually.

Self-host VPS sizing

Default is Hetzner Cloud (the best price/performance for n8n at the moment). Tier picked by load:

Equivalent DigitalOcean droplets ($12 / $24 / $48) and AWS t3.small / t3.medium / t3.large on-demand are roughly 2-3x more expensive at the same RAM tier. Pricing for those is published in the same panels.

Ops time

Self-hosting is not free even when the VPS is. The calculator adds a rule-of-thumb ops cost: 1 hour/month for a single instance, 3 hours/month once queue mode is on (Postgres updates, Redis snapshots, worker restarts, n8n version upgrades). Multiplied by your hourly rate.

Queue mode trigger

Auto mode flips queue mode on when executions exceed 50,000/month, when 5+ workflows hit concurrent triggers, or when a single execution would otherwise block others. Below that, single-instance is fine and meaningfully cheaper to run. You can override the toggle if your workload says otherwise.

Break-even

The "self-host wins above N executions/month" line is the volume at which monthly self-host cost (VPS + add-ons + ops time at your rate) crosses the cheapest Cloud tier that would fit. Drop your DevOps hourly rate, the line moves down. Raise it, the line moves up.

What this calculator does not include

FAQ

Is n8n Cloud always more expensive at high volume?

At very high volume (above the Pro tier), Cloud moves to Enterprise pricing, which is negotiated case-by-case. Self-hosting on a single Hetzner CX42 with queue mode comfortably handles hundreds of thousands of executions per month at <$25 of compute, plus your ops time. For most teams above ~30k executions/month, self-host is cheaper even after charging realistic ops time.

Why does the calculator not let me enter every Cloud tier manually?

Because Cloud caps execution count rather than charging per-execution overage on the self-serve tiers. You either fit in a tier or move up. The calculator picks the cheapest fit, which is what n8n's billing actually does.

Why Hetzner as the default VPS?

Hetzner Cloud is the price/performance leader for n8n right now - roughly half the cost of DigitalOcean for the same RAM tier, and a third of AWS on-demand. The calculator shows DO and AWS equivalents in the methodology section so you can re-price against whatever you actually run.

What about n8n Embed and white-label pricing?

Out of scope for this calculator. Embed and Enterprise terms are negotiated, not posted. This tool covers the public Starter, Pro, and self-host paths.

Can I share my numbers with a teammate?

Yes - the URL updates as you change inputs. Copy the address bar after you fill the form and the link reproduces your scenario for whoever opens it. There is also a "Copy as Markdown" button that produces a table you can paste into Slack, Notion, or a Reddit reply.

Want us to run the migration?

AutomateLab moves teams between n8n Cloud and self-hosted (in either direction) as a fixed-scope project. Two-to-four weeks, runbook included, your repo.

Tell us about your setup