GLIVENTRequest a review
Automation Tools and Platforms

What Does n8n Actually Cost — Cloud vs Self-Hosted?

Last updated 22 July 2026 · 6 min read

Direct Answer

n8n's cloud plans are priced by workflow executions per month (a workflow "executes" once each time it runs end to end, regardless of how many steps it contains), with tiers scaling up in execution allowance and price. Self-hosting n8n costs nothing for the software itself — it's source-available (fair-code licensed), free to self-host for internal business use — but isn't free in practice: you pay for server hosting, and for someone's time to set it up, keep it updated, and fix it when it breaks. Which is cheaper depends on your execution volume and whether you already have the technical capacity to run infrastructure; check n8n's current pricing page before budgeting, since tiers and rates change.

Detailed Explanation

n8n prices itself differently from Zapier, Make, and Power Automate, which trips up anyone comparing the four on cost alone. Where Zapier counts individual action steps (see what counts as a task in Zapier), n8n's cloud plans bill by workflow executions — one execution per end-to-end run of a workflow, no matter how many steps or nodes it contains. That makes an execution-heavy but simple workflow cost the same as a step-heavy complex one running the same number of times, which is the opposite of how Zapier's per-action counting works.

n8n's cloud tiers scale by monthly execution allowance, with higher tiers unlocking more executions, more active workflows, and features like shared team environments. Exact tier names, execution caps, and prices change often enough that quoting specific figures here would go stale — check n8n's own pricing page for current numbers before budgeting.

The other half of the decision is n8n's self-hosted option, which doesn't exist for Zapier, Make, or Power Automate in the same form: because n8n is source-available and free to self-host for internal use, you can run it on your own server (or a cloud VM you control) instead of paying for its hosted cloud plans at all.

Cloud Pricing: What You're Actually Paying For

n8n's cloud plans are the SaaS option — you sign up, don't manage any infrastructure, and pay a monthly or annual fee tied to an execution allowance. The practical planning questions are the same ones that matter for any usage-based automation platform:

  • Estimate your execution volume realistically before picking a tier. A workflow that runs once per new lead behaves very differently, cost-wise, from one polling a mailbox every few minutes regardless of whether new mail arrived.
  • A scheduled or polling trigger can generate executions even when nothing "useful" happens, depending on how the workflow and trigger are configured — check whether your trigger only executes on new data or on every check, since that materially changes your monthly execution count.
  • Team and environment features are usually gated to higher tiers, not just execution volume — if you need multiple people editing workflows or separate staging/production environments, factor that into which tier you actually need, not just raw execution count.

Self-Hosted: The Real Cost Is Time, Not Software

Self-hosting n8n means installing and running the software yourself, typically via Docker on a cloud VM (AWS, DigitalOcean, or similar) or on-premises infrastructure. The software license costs nothing. What it actually costs:

  • Server hosting. A small VM capable of running n8n reliably typically costs a modest monthly amount — inexpensive on its own, but not zero, and it scales up if your workflow volume or data processing needs grow.
  • Setup time. Getting n8n installed, configured, and connected to your business systems takes real hours from someone with at least moderate technical comfort — this isn't a five-minute signup like a cloud plan.
  • Ongoing maintenance. Software updates, security patches, backups, and monitoring uptime are now your responsibility, not a vendor's. n8n releases updates regularly; skipping them for too long can leave you on an unsupported version when you eventually do need to upgrade.
  • Troubleshooting when it breaks. If self-hosted n8n goes down, there's no vendor status page or support ticket queue to lean on the way there is with a SaaS platform — see what happens to your automations during a Zapier, Make, or Power Automate outage for how that vendor-outage risk compares; self-hosting trades that risk for owning your own uptime instead.

For a business with in-house technical capacity (an IT-comfortable employee or a developer on staff), self-hosting can be genuinely cheaper at meaningful execution volume, since there's no per-execution billing ceiling. For a business without that capacity, the time cost of standing up and maintaining a server often erases the savings versus simply paying for a cloud plan.

Which Fits Your Business

Work through this in order:

  1. Do you have someone technical enough to comfortably run a Docker container and apply updates? If not, self-hosting isn't realistically free — it's a hidden ongoing cost in that person's time, or a project you'd need to hire out. Choose n8n cloud, or reconsider whether Zapier, Make, or Power Automate fit better for a non-technical team (see Zapier vs Make vs n8n vs Power Automate).
  2. Is your execution volume high enough that cloud tier pricing would get expensive? High-volume, technically capable businesses are where self-hosting's economics work best — no per-execution billing ceiling once the server is paid for.
  3. Does data residency or control matter — do workflows touch data that can't leave infrastructure you control? Self-hosting keeps workflow data on your own servers, which matters for some compliance or contractual requirements the same way it does for other automation-data decisions (see how do you securely manage credentials and connected accounts in an automation platform).
  4. None of the above tip it clearly? Start on n8n cloud. It's the lower-commitment option, and workflows built there carry over if you later decide self-hosting makes sense.

Things to Consider

  • Pricing details change — this page describes the model, not exact figures. Confirm current tier names, execution allowances, and prices on n8n's own pricing page before committing a budget.
  • Execution-based pricing rewards workflow consolidation differently than action-based pricing does. Because n8n counts one execution regardless of step count, combining several small workflows into fewer, more comprehensive ones can reduce your execution count in a way that wouldn't help on a per-action platform like Zapier.
  • A cheap self-hosted server is not the same as a reliable one. Running n8n on the smallest possible VM to save money can create the exact silent-failure risk covered in how do you stop an automation from failing silently — under-provisioned infrastructure timing out or crashing under load.

Common Mistakes

  • Comparing n8n's headline price directly against Zapier's or Make's without adjusting for the different pricing model. Execution-based and action-based pricing don't translate one-to-one; model your actual usage on each platform rather than comparing sticker prices.
  • Treating self-hosting as free because the software is. The software license costs nothing; the hosting and — usually larger — the maintenance-time cost are real and ongoing.
  • Self-hosting without a named person responsible for keeping it updated. An unmaintained self-hosted instance is a security and reliability risk, not a cost-saving win, the same trade-off flagged for self-hosted platforms generally in Zapier vs Make vs n8n vs Power Automate.
  • Not checking whether a polling trigger executes on every check versus only on new data. This single configuration detail can be the difference between a low and a surprisingly high monthly execution count on a cloud plan.

Frequently Asked Questions

Does an n8n workflow with many steps cost more than a simple one?
Not directly. n8n's cloud pricing counts executions — one execution per end-to-end workflow run — not actions or steps within it, which is a different model from Zapier and Make, where cost scales with the number of action steps a workflow contains. A ten-step n8n workflow and a two-step one both use one execution per run; what drives cost is how often the workflow runs, not how complex it is.
Is self-hosted n8n really free?
The software is free to self-host for internal business use (n8n is source-available under a fair-code license, not traditional open source), but running it isn't a zero-cost decision. You still pay for a server (even a small one has a monthly hosting bill), and — the larger cost for most small businesses — someone's time to install it, apply updates, monitor uptime, and troubleshoot it when something goes wrong. For a business without existing technical capacity, that time cost often exceeds what a cloud plan would have charged.
Can you switch between n8n cloud and self-hosted later?
Yes — workflows built in n8n are portable between its cloud and self-hosted versions since both run the same underlying software, which is a meaningful advantage over switching between entirely different platforms (see how do you migrate an automation workflow from Zapier to Make for how much harder that cross-platform version is). Moving from cloud to self-hosted, or the reverse, is realistic if your needs change, though you should still budget time to test the migration rather than assuming it's instant.

References

Related Questions

Zapier vs Make vs n8n vs Power Automate — How Do You Choose?

Zapier is easiest to start with, Make handles complex logic better, n8n suits technical teams, and Power Automate fits Microsoft 365 businesses.

What Counts as a Task in Zapier (and Why Is Your Usage Burning So Fast)?

A Zapier task is each successful action a Zap runs — multi-step Zaps use one task per action, so usage burns faster than most people expect. Here's why.

How Does Make's Operations-Based Pricing Work (and How Do You Avoid Burning Through Your Plan)?

Make counts operations per module run, including many trigger checks — a scenario with several modules or frequent polling burns operations fast.

How Do You Stop an Automation From Failing Silently?

Every platform needs failure alerts turned on manually — Zapier, Make, Power Automate, and n8n included — or a broken flow runs silently.

How Are AI Agents Priced, and How Do You Cap the Cost?

AI agent platforms are priced by seat, by task, by usage credits, or by outcome — and the model matters as much as the sticker price. Here's how to cap spend.

When Is It Worth Switching Automation Platforms (and When Should You Stay)?

A rising bill alone isn't a reason to switch automation platforms. How to tell a real structural problem from billing frustration, and what switching costs.