n8n Pricing in 2026: Every Tier, True Cost, and When Self-Hosting Wins

TL;DR: n8n Cloud starts at €20/month and ends at €667/month before custom Enterprise. The cloud pricing model is generous compared to Zapier or Make for most workflows because n8n counts executions per workflow run, not per step. But the moment your workflows poll frequently or your team grows past three contributors, self-hosted n8n on a $5 VPS becomes the rational choice. Below: every tier, what each one actually buys you, the four hidden costs the pricing page does not mention, and the cost comparison with Zapier and Make at real team sizes.

What you actually pay

n8n Cloud has four tiers. All require annual billing for the listed prices (monthly billing is roughly 20% more expensive).

All plans include unlimited users, unlimited workflows, and unlimited workflow steps. The constraint is execution count, where one execution equals one full workflow run regardless of how many nodes it touches.

The Business plan adds a critical capability the lower tiers do not: self-hosted deployment with a license key. Cloud-hosted Business is the same price; the choice exists for teams that need data residency, VPC peering, or air-gapped environments.

The execution counter is the part people misread

Cloud pricing is anchored on executions. The trap is that polling workflows burn through quotas faster than most users predict.

A workflow that checks Gmail every 10 minutes runs 4,320 times per month. A Slack monitor checking every 5 minutes runs 8,640 times. Both blow past the Starter plan's 2,500 cap on day one. Webhook-triggered workflows are cheap; polling-triggered workflows are not.

The fix: prefer webhooks over polling wherever the source platform supports them (Slack, Stripe, Linear, GitHub all do). For sources that only support polling, batch your checks every 30+ minutes unless real-time matters for the use case. This single change typically cuts execution counts by 70%+.

The hidden costs nobody publishes

Four costs hit accounts on Pro and above that the pricing page does not flag clearly.

1. Overage fees on Business. Once you exceed your 40,000 monthly execution quota, n8n charges €4,000 for every additional bucket of 300,000 executions. That works out to roughly €0.013 per execution, which is reasonable until you triple your volume mid-month and end up with an unexpected €8,000 invoice. The fix is either upgrading proactively or moving the offending workflows to self-hosted.

2. Annual-only pricing on the cheap tiers. Monthly billing exists, but the prices are 20% higher and not displayed on the public pricing page. If you want to test n8n Cloud for a quarter without committing 12 months upfront, that flexibility costs roughly €4/month extra on Starter and €10/month on Pro.

3. AI credits run out fast. The Starter plan includes 50 AI credits per month, which sounds reasonable until you remember a single AI node call against GPT-4o or Claude Sonnet 4.6 burns through one credit. A workflow that drafts replies for 50 inbound emails consumes the entire monthly allocation. Pro doubles this to 150 credits, still thin for any AI-heavy automation.

4. Workflow history is short on lower tiers. Starter retains one day of execution history. When a workflow fails Friday night and you discover it Monday morning, the failure data is gone. Pro retains five days, which covers a weekend. Business retains 30 days. For mission-critical workflows, anything less than Pro is a debugging liability.

When self-hosted beats Cloud

n8n's Community Edition is fully open-source under the Sustainable Use License. You can run it on any server with Docker, the binary is identical to the cloud offering, and there is no execution cap.

The total cost of self-hosting breaks down like this:

For a team running fewer than 50,000 executions per month and comfortable with Docker, raw VPS self-hosting saves roughly €40-50/month versus Cloud Pro. For teams above 50,000 executions, the savings versus Cloud Business become substantial (€600+/month).

The catch: self-hosting loses the AI credits, the managed updates, and the support SLA. Most importantly, advanced features (SSO, audit logs, external secrets, version control on workflows) require a Business or Enterprise license key even on self-hosted deployments. That license starts at €667/month, which erases the cost savings unless the workload genuinely needs those features.

n8n vs Zapier vs Make at real team sizes

Pricing comparisons between automation platforms are misleading because they count units differently. Zapier counts tasks (one task = one action step). Make counts operations (one operation = one module call, including condition checks). n8n counts executions (one execution = one full workflow run, regardless of step count).

A workflow with 8 steps that runs 1,000 times per month is:

• Zapier: 8,000 tasks → Professional tier (~$50/mo for 10K tasks)
• Make: 8,000 operations → Core tier ($9/mo for 10K operations)
• n8n: 1,000 executions → Starter tier (€20/mo)

At low volume, Make wins on price. At higher volume, n8n's execution-based counting becomes the cheapest option by a significant margin.

For teams above 10,000 monthly runs that can self-host, n8n is the cheapest option by an order of magnitude. For teams that need a managed experience, Make is cheapest at low volume and n8n is cheapest at high volume.

Which plan you actually need

Solo founders, side projects, indie developers: Self-host on a $5/month Hetzner VPS. The 2-hour setup pays for itself within the first month. If Docker scares you, Elestio or Pipedream managed hosting at $15/month is a reasonable middle ground.

Small teams (2-10 people), simple automations: Cloud Pro at €50/month. The 10K execution quota covers most teams' workflows, and the 20 concurrent runs handle bursts. Five days of history is enough for weekend-fail recovery.

Mid-size teams (10-50 people) or AI-heavy workflows: Cloud Business at €667/month if you need 40K executions and 30-day history. Otherwise, self-host with a Business license for the same €667/month but cheaper infrastructure at higher volume.

Enterprise / regulated industries: Enterprise tier. The negotiable items are execution volume, support SLA, and whether n8n hosts or you self-host. The published price is a starting point; teams above 250K monthly executions typically negotiate 20-30% discounts on annual contracts.

Skip n8n if...

The execution model is wrong for two use cases. Skip n8n if your workflows are heavily polling-based against APIs without webhook support (Zapier's task counting can be cheaper). Skip n8n if your team has zero DevOps capacity and zero budget for managed hosting (the Cloud Pro tier at €50/month is competitive but not the cheapest option for small workloads).

For everything else, especially AI-heavy automation, internal tools, and high-volume webhook workflows, n8n's pricing is the best in the category.

FAQ

Is n8n's Community Edition really free? Yes, fully free under the Sustainable Use License. The only restriction: you cannot resell n8n itself as a competing automation platform. Internal use, client work, and commercial workflows are all permitted.

Can I move from Cloud to self-hosted later? Yes. Workflows export as JSON and import cleanly into a self-hosted instance. Credentials need to be re-entered (not exportable for security reasons). Most teams migrate over a weekend.

What's the cheapest way to run n8n at scale? Self-hosted on a $5-15/month VPS with PostgreSQL backup, automated by a $0 GitHub Actions cron for nightly snapshots. Total monthly cost: under $20 for unlimited executions, including backup infrastructure.

Are there annual prepayment discounts? Yes. All Cloud tiers are 17-20% cheaper on annual billing versus monthly. Enterprise contracts typically add another 10-20% discount for multi-year commitments.

Does the Business plan include a self-hosted license? Yes. Business and Enterprise both include the EE license key required to unlock SSO, audit logs, external secrets, and worker mode on self-hosted deployments. The license is tied to your n8n account, not the deployment.

Bottom line

For most teams making the n8n vs Make vs Zapier decision in 2026, n8n is cheaper at scale and more powerful for AI-heavy workflows. The Cloud tiers are fairly priced if you want managed hosting; the Community Edition on a $5 VPS is the unbeatable option for anyone comfortable with Docker. The trap to avoid is letting polling-based workflows quietly burn through execution quotas. Audit your trigger types, prefer webhooks, and the pricing math always works in your favor.