n8n Cloud vs Self-Hosted on Hostinger (2026): The Real Math

The decision table

n8n bills Cloud plans in EUR — there is no USD price list on n8n.io, so the euro figures above are the official prices (monthly billing; annual billing runs about 17% cheaper). Your card provider converts at checkout. Business tier (€800/mo, 40,000 executions) exists for teams needing SSO and governance; it is overkill for almost every reader of this page, so the comparison below focuses on Starter, Pro, and self-hosting.

What do you actually get on n8n Cloud?

n8n Cloud is n8n-the-product without n8n-the-server. You sign up in the browser, and within minutes you are building workflows in the same visual editor the self-hosted version uses. n8n handles hosting, updates, security patches, and backups of your instance; you never SSH anywhere or think about Docker. The pricing unit is the execution: one complete run of a workflow, no matter how many steps it contains. A 30-node workflow that enriches a lead, calls three APIs, and writes to two systems counts as one execution — which is why n8n Cloud is dramatically cheaper than per-operation platforms like Zapier for wide, complex workflows, and comparatively worse for thousands of trivial two-step runs.

The tiers are simple. Starter at €24/month gives you 2,500 executions a month (roughly 83 a day) with 5 concurrent runs — enough for a freelancer’s first production automations. Pro at €60/month gives you 10,000 executions with 20 concurrent runs, plus workflow history, global variables, and admin roles for small teams. Both include unlimited active workflows and unlimited users, and both come with a 14-day free trial that needs no credit card, so you can measure your real execution volume before committing. That measurement step matters more than any advice on this page: most people wildly misestimate how many executions they run.

The catch is the ceiling. When you exceed your plan’s execution allowance, n8n Cloud pauses your workflows until the next billing cycle — there is no pay-as-you-go overage to smooth a spike. A viral week, a runaway AI agent loop, or a misconfigured polling trigger can therefore silence your automations at exactly the worst moment. Heavy users either buy headroom they do not need most months or start eyeing the self-hosted column of the table above.

What does self-hosting on a Hostinger VPS really cost?

Self-hosting has two cost lines: the software, which is free, and the server, which is not. The n8n Community Edition — the self-hosted version — costs nothing to license and imposes no execution limit; your ceiling is whatever the server can handle. On a Hostinger VPS KVM 2 (2 vCPU, 8 GB RAM, 100 GB NVMe), that ceiling is high: thousands of executions a day without breaking a sweat, and Hostinger’s own tutorial calls KVM 2 “more than enough” to get started. Deployment is genuinely minutes via the one-click “Ubuntu 24.04 with n8n” template, which ships n8n preinstalled in Docker.

The server price is the live card below — intro price, term, and renewal, checked October 9, 2026. That is the entire software bill: no per-execution meter, no per-seat tax, no tier cliff when you grow from 9,000 to 11,000 executions. The second cost line is your time: you own the operating system, n8n updates, backups, and uptime. Budget an hour or two a month once it is stable, more in the first month while you harden it (HTTPS from day one, strong owner password, and — the one everyone forgets — your N8N_ENCRYPTION_KEY stored safely off the server, because without it your stored credentials are unreadable even from a backup).

Our full setup walkthrough — template choice, sizing by workload, queue mode for heavy setups, and the failure modes nobody warns you about (502s, out-of-memory kills, SQLite vs Postgres) — lives in our n8n hosting guide. If you are new to servers, read it before you buy anything; it will tell you honestly whether self-hosting fits your patience.

Where is the break-even point?

Here is the math, with nothing hidden. One year of n8n Cloud Starter costs €288 (€24 × 12); one year of Pro costs €720. The self-hosted column costs the 24-month VPS total in the box below, divided by two for a yearly figure — do that division with the live numbers, because intro pricing moves and we will not print a stale figure here. At typical KVM 2 pricing, a single year of self-hosting lands well under €288, which means the break-even against Starter sits somewhere in the first few months: after that, every additional month is money Cloud would have charged and the VPS does not.

Against Pro the comparison is almost unfair: €720 a year buys roughly five to seven years of VPS time at common intro rates, and the VPS has no 10,000-execution ceiling while Pro pauses your workflows at 10,001. The honest counterweight is your time. If an hour of your month is worth more than the price gap, and your volume fits comfortably in Starter’s 2,500 executions, Cloud is genuinely the cheaper option — the server you do not administer is the cheapest server of all. Past ~10,000 executions a month, or the moment a paused workflow would cost you real money, self-hosting wins on both price and peace of mind.

One more scenario people miss: the growing team. n8n Cloud’s jump from Pro (€60) to Business (€800) is a 13x price cliff for 4x the executions, and it exists mostly to gate SSO, Git, and environments. A team outgrowing Pro faces either that cliff or self-hosting — and self-hosted n8n with a Postgres database handles team workloads on the same KVM 2 without blinking. If your roadmap has “more automations” anywhere on it, price the destination, not just the starting tier.

What happens when you hit the execution limit?

This deserves its own section because it is the least understood part of n8n Cloud. There is no overage billing: when your plan’s executions run out, n8n pauses all workflows until the next billing cycle starts. No warning email buys you extra runs; the automations simply stop. For a nightly report workflow that is an annoyance. For order processing, lead routing, or anything customer-facing, it is an outage you did not schedule.

Two things make this bite harder than it sounds. First, sub-workflows called from a parent workflow generally count as their own executions, so a tidy, modular automation architecture consumes executions faster than its workflow count suggests. Second, AI agent loops are execution-hungry by nature: an agent that iterates, retries, and calls sub-workflows can burn through a month’s allowance in days. If either describes your setup, either buy far more headroom than your current usage suggests or self-host, where the only limit is server capacity.

Self-hosting inverts the risk: nothing pauses, ever. The failure mode moves from “paused by billing” to “slow under extreme load,” which you fix by scaling the server — KVM 4 when queue mode with separate workers enters the picture. Monitor execution counts either way; n8n’s dashboard shows them, and the 14-day Cloud trial is the perfect place to measure your real volume before you choose.

Who should pick n8n Cloud?

Pick n8n Cloud if your execution volume is modest and your time is expensive. Freelancers automating client onboarding, small teams with a handful of scheduled workflows, and anyone who wants automation running this afternoon without learning Docker are the textbook fit. The 14-day free trial with no credit card makes this a low-risk experiment: build your real workflows, watch the execution counter for two weeks, and extrapolate. If you land well under 2,500 a month with room to spare, Starter at €24 is rational money for never thinking about servers.

Cloud is also the right answer when someone else’s uptime is your problem and you have no one to page: solo founders, non-technical teams, and businesses where automation is important but not the product. n8n’s team handles the 3am problems. Just go in with eyes open about the ceiling — set up execution alerts, keep 30% headroom, and have a plan (upgrade or migrate) for the month you outgrow the tier.

Who should self-host instead?

Self-host if any of these are true: you run (or will run) more than ~10,000 executions a month; a paused workflow would cost you money or customers; you handle data that should not leave your infrastructure; you want AI agents iterating freely without a meter running; or you simply enjoy owning your stack. Developers, agencies running client automations, and data-conscious EU businesses cluster here for good reason — unlimited executions plus full data control at a flat server price is a combination Cloud cannot match at any tier.

The prerequisite is honest self-assessment about server administration. You will keep the OS patched, n8n updated, backups verified, and the encryption key safe. The one-click template removes the installation pain but not the ownership. If that paragraph felt like a chore list, take the Cloud trial first — you can always migrate to a VPS later, when the execution bill tells you it is time. Our n8n hosting guide covers the migration-relevant details (Postgres vs SQLite, queue mode) so a later move does not surprise you.

Pros and cons

FAQ

Related