n8n pricing has two honest answers, and which one applies to you depends entirely on who maintains it. Self-hosting the community edition carries no license fee: you run it on your own server under n8n's fair-code Sustainable Use License and pay only for infrastructure and your time. n8n Cloud, the managed version, starts at 20 euros per month billed annually as of July 2026, and its billing unit, the execution, is the single most important thing to understand before comparing it with anything else. This guide covers the plans, the execution math, the true cost of the "free" option, and the scenarios where each path wins. For the full head-to-head with its biggest rival, see Zapier vs n8n.
The short answer
n8n costs one of three things, and each corresponds to a different buyer. All numbers here are as published on n8n's pricing page in mid-2026; verify current figures there before you commit.
- Zero license fee, self-hosted. The community edition on your own server. You pay for a small VPS and, more meaningfully, for the technical time to run it. Best for developers and teams with data-control requirements.
- 20 euros per month billed annually, n8n Cloud Starter. Managed hosting with 2,500 workflow executions included. Best for technical builders who want n8n's power without the server. Annual billing saves 17 percent versus monthly.
- 50 euros per month for Pro, more above it. 10,000 executions, higher concurrency, team features; Business sits at 667 euros per month for scale and SSO, and Enterprise is custom. Best when n8n becomes core infrastructure.
What n8n never charges for is step count. That is the structural difference from per-task competitors, and it is worth a section of its own.
n8n Cloud plans at a glance
n8n Cloud's ladder is short and legible. The tiers differ mainly in execution allowance, concurrency, and collaboration features rather than in which nodes you can use, which keeps the comparison honest.
| Option | Price (July 2026, billed annually) | What you get | Best for |
|---|---|---|---|
| Community (self-hosted) | No license fee | Full workflow engine on your infrastructure, community support | Engineers, data-control requirements |
| Cloud Starter | 20 EUR/month | 2,500 executions, 5 concurrent, unlimited users | Technical builders without ops appetite |
| Cloud Pro | 50 EUR/month | 10,000 executions, 20 concurrent, admin roles, workflow history | Teams running n8n as core plumbing |
| Business | 667 EUR/month | 40,000 executions, SSO/SAML, environments, Git version control | Companies scaling n8n org-wide |
| Enterprise | Custom | Custom executions, 200+ concurrency, dedicated support, log streaming | Compliance-heavy organizations |
Three notes worth knowing before you pick a tier. First, allowances are about executions, not steps, so estimate your run counts rather than your workflow complexity. Second, the prices above are annual billing, which saves 17 percent versus paying monthly, and startups under 20 employees can qualify for half off Business. Third, the community edition is source-available fair-code, not OSI open source; for internal business automation that distinction rarely matters, but it does restrict reselling n8n as a service.
How execution billing works, and why it matters
An execution is one complete run of one workflow, whether that workflow has three nodes or thirty. Every trigger fire that starts a run consumes exactly one execution from your allowance. Steps, branches, and loops inside the run are free.
This is the mirror image of per-task billing, where every action step in every run is metered. The consequence: as your automations grow more sophisticated, n8n's effective price per unit of work falls, while per-task tools charge you for the added sophistication. A workflow you enrich from five steps to fifteen costs exactly the same per run on n8n and three times as much on a per-task meter. We walk through the full comparison, including where each model wins, in Zapier pricing explained and the head-to-head Zapier vs n8n.
The real cost of self-hosting "free" n8n
License-free is not cost-free, and the teams happiest with self-hosted n8n are the ones that priced this honestly on day one. The bill has three parts, only one of which appears on an invoice.
- Infrastructure. A small VPS runs n8n comfortably for most workloads; think single-digit dollars per month at hobby scale, more with volume, queues, and a proper database.
- Setup and upkeep. Installing, configuring credentials, applying updates, managing backups, and securing the instance. Plan on a few hours of technical time per month as the steady state.
- Failure handling. When a workflow silently stops at 2 a.m., you are the on-call. This cost is invisible until the first incident and decisive after it.
The rule of thumb we see hold up: self-hosting pays off when a technically comfortable person already exists on the team and would enjoy owning it. Hiring or contracting that capability just to save a 20-euro subscription is a losing trade. If lock-in is what draws you to self-hosting, note that portability exists at the agent layer too; see AI agent platforms with no vendor lock-in.
Worked cost scenarios
Start with the arithmetic that separates the two billing models. The table takes one five-step workflow at three monthly volumes and counts the billing units each meter consumes; the units are the whole story, because plan prices then scale with them.
| Runs per month (5-step workflow) | n8n executions consumed | Zapier tasks consumed | Where that lands you |
|---|---|---|---|
| 1,000 | 1,000 | ~5,000 | n8n Starter (2,500 allowance) handles it; Zapier needs a mid task tier |
| 5,000 | 5,000 | ~25,000 | n8n Pro (10,000) is comfortable; Zapier needs a high-volume tier |
| 10,000 | 10,000 | ~50,000 | n8n Pro exactly fits; the per-task bill is now the dominant line item |
When we priced a typical dozen-workflow ops stack both ways, the finding was consistent: the meter, not the headline price, decided the bill once workflows passed three or four steps. Three realistic shapes show where each option lands.
- Light automation: five simple workflows, ~300 total runs a month. Starter's 2,500-execution allowance swallows this without noticing. Self-hosting saves nothing after you count an hour of maintenance. Verdict: Cloud, or honestly any tool.
- Heavy plumbing: a dozen multi-step workflows, ~5,000 runs a month. This is where execution billing shines: 5,000 executions regardless of step count, versus tens of thousands of billing units on a per-task meter. Verdict: n8n Cloud Pro at 50 euros, or self-host if you have the engineer.
- Variable, judgment-heavy work: research, triage, reporting. Run counts are low but every run is different, and the cost that dominates is the human time spent designing and repairing flowcharts for work that keeps changing shape. Verdict: this is the scenario where workflow tools, at any price, are the wrong shape; see the next section.
When n8n is a bargain, and when price is beside the point
n8n is arguably the best-value workflow engine in the category when three things are true: someone technical owns it, workflows are multi-step and high-volume, and the work is stable enough that a flowchart built once keeps paying off. Under those conditions, execution billing plus optional self-hosting is hard to beat, which is exactly why n8n tops so many lists in our n8n alternatives roundup rather than appearing on them.
The price stops mattering when the premise fails: when nobody wants to own workflow logic, or the work varies too much to diagram. A cheap engine you must operate is not cheap if operating it is the part you were trying to escape. That boundary, workflows you design versus outcomes you describe, is the real dividing line in this market, and it is why AI agent vs workflow automation is a more useful comparison than any two workflow tools against each other. Teams weighing that switch can start with when to switch from Zapier to an AI agent; the logic applies to n8n stacks equally.
The outcome-based alternative
If what you actually want is the result, the weekly report assembled, the inbox triaged, the leads chased, without anyone building or maintaining the machinery, the relevant price is not per execution but per outcome. On Gravity, you describe the task in plain words, the right expert-built agent runs it, and a human approves the moments that matter. Pricing is a subscription: a free tier with one agent at 0 dollars, then plans from 20 dollars per month that include 20 dollars of usage, with extra usage purchasable as you grow. The comparison with n8n specifically is covered in Gravity vs n8n, and the wider budget landscape in cheapest AI agent platforms and AI agent pricing explained.
Frequently asked questions
Is n8n free to use?
The self-hosted community edition has no license fee under n8n's Sustainable Use License, and it allows internal business use. Your real costs are the server it runs on and the hours spent installing, updating, backing up, and securing it. n8n Cloud, the hosted version, is a paid subscription.
How much does n8n Cloud cost per month?
As of July 2026, n8n Cloud's Starter plan is 20 euros per month billed annually with 2,500 workflow executions included, Pro is 50 euros per month with 10,000 executions, and Business is 667 euros per month with 40,000. Enterprise pricing is custom, and annual billing saves 17 percent versus monthly. Check n8n's pricing page for current numbers before committing.
What counts as an execution in n8n?
One complete run of one workflow counts as one execution, no matter how many nodes it passes through. A thirty-node workflow and a three-node workflow each consume a single execution per run, which is why n8n gets comparatively cheaper as your workflows get more complex.
Is n8n cheaper than Zapier?
Usually, once workflows have several steps or real volume. Zapier bills per task, meaning every action step in every run; n8n bills per execution, meaning one count per run. A five-step workflow running 1,000 times a month consumes roughly 5,000 Zapier tasks but only 1,000 n8n executions. At one or two steps and low volume, the difference mostly disappears.
What are the hidden costs of self-hosting n8n?
The server bill is the small part. The real costs are operational: applying updates, managing credentials securely, monitoring uptime, handling backups, and debugging when a workflow silently stops. Budget a few hours of technical time per month, and more when something breaks at a bad moment.
What if I do not want to run or maintain workflows at all?
Then the pricing comparison that matters is workflows versus outcomes. An AI agent platform takes a plain-language description of the result you want and handles the steps itself. Gravity's model is a free tier with one agent, then subscription plans from 20 dollars per month that include 20 dollars of usage.
Sources
- n8n, "Pricing", n8n.io/pricing, backs all Cloud plan figures cited (Starter 20 EUR/month annual with 2,500 executions, Pro 50 EUR with 10,000, Business 667 EUR with 40,000, 17 percent annual saving, startup discount), verified live 2026-07-24; current numbers live there.
- n8n, "Sustainable Use License", docs.n8n.io, backs the fair-code license description, the internal-business-use allowance, and the resale restriction.
- Zapier, "Plans & Pricing", zapier.com/pricing, backs the per-task billing model used in the comparison math.
- Gravity, "How it works", gravity.fast, backs the outcome-based description and Gravity's subscription pricing.