n8n Pricing: Cloud Plans, Executions and Self-Hosting

n8n Cloud costs 20 euros a month for Starter (2.5K workflow executions) and 50 euros for Pro (10K), billed annually. The Business plan is 667 euros a month for 40K executions and runs on your own servers. Community, the self-hosted edition, is free. You pay per workflow run, not per user or step. Prices checked 2026-10-31.

Prices checked 2026-10-31

Source: n8n.io/pricing and the n8n docs. Prices are euros per month when billed annually, which the page says saves 17 percent against monthly. The page we fetched showed the annual figures only. Its Monthly toggle shows monthly prices, which we did not capture.

All plans include unlimited users, unlimited workflows and every integration. n8n says pricing is "based on monthly workflow executions, regardless of complexity." For the tool itself, see n8n and the build guide on n8n AI agent tutorial.

What an execution is

An execution is one run of an entire workflow. It does not matter how many steps the workflow has or how much data passes through it. n8n contrasts this with tools that bill per task or operation, and says a single workflow can do in one execution what would take thousands of operations elsewhere. That claim is n8n's own, and we did not test it.

The page adds a billing principle: "If it doesn't run, you don't pay". You pay only when a workflow runs from start to finish.

Estimating your executions

n8n gives three rules of thumb.

A worked example. Five scheduled workflows that run every 15 minutes each use about 2,880 executions a month apiece, so 14,400 in total. That is over Pro's 10K. The same five at hourly intervals use about 3,600, which fits Pro but not Starter. These are our sums from n8n's counting rules, not a measurement of any real instance.

Production execution counts appear in the Insights dashboard. Business subscribers get a weekly email with counts across every instance the license key is applied to.

What each plan includes

Some limits differ by plan in the comparison table: saved executions are 2.5K on Starter, 25K on Pro and 50K on Enterprise. Maximum execution duration is 5 minutes on Starter and 40 minutes on Pro and Enterprise. Execution log retention is 7 days on Starter, 30 on Pro and unlimited on Enterprise. The page's concurrency figures conflict a little: the plan cards say Pro runs up to 50 concurrent executions, while the comparison table says 20 / 50. Pro comes in more than one size, so read the limit as 20 or 50 depending on the size you pick.

Community edition and self-hosting

The self-hosted Community edition is free and, per n8n's docs, "includes almost the complete feature set." It leaves out these features, which sit on paid plans:

• custom variables
• environments
• external secrets
• external storage for binary data
• log streaming
• multi-main mode
• projects
• SSO (SAML, LDAP)
• sharing of workflows and credentials
• version control using Git

Without sharing, only the instance owner and the person who created a workflow or credential can access it. That is the main practical limit for teams.

Self-hosting has three tiers in n8n's own words. Community is free and runs without a license key. Registered Community is the same edition with extra features unlocked by registering your email, still free. Business and Enterprise are paid, and subscribing gives you a license key that unlocks that plan's features. One key works on unlimited instances, with combined usage counted toward your quota. The key must ping n8n's license server daily, and the ping includes the number of production executions.

You pay for hosting on top. We have not priced a server here. See our self-host LLM guide and the host entries for RunPod and Modal for infrastructure if you run local models alongside.

Business plan: the one that surprises people

Business is 667 euros a month for 40K executions and is for self-hosted use only. The page says it is "for companies with < 100 employees needing collaboration and scale." Cloud-hosted Business is not available yet, and n8n has a waitlist. If you want hosted and collaborative, the page offers Pro or Enterprise.

Business is a self-serve plan with forum support only. Dedicated support with an SLA is Enterprise only. Payment can be by invoice or wire transfer for annual Business and Enterprise. Monthly Business can be cancelled any time at the close of the billing cycle, annual ends at the annual term, and n8n does not offer pro-rata refunds.

If you exceed your Business quota, workflows keep running, and overage charges may apply unless you move up a usage tier. The sales team contacts annual subscribers near 80 percent of quota.

For small companies, n8n has a Startup plan: under 20 employees may qualify for 50 percent off Business. That would be about 333 euros a month for the same 40K executions on annual billing, by our arithmetic from the stated discount. We did not read the Startup plan's eligibility terms.

Free trial

A Cloud trial gives Pro-level features with trial limits: 1,000 executions, 5 concurrent executions, a 180-second execution timeout and 1 team project. It includes 800 Assistant credits plus 100 AI credits. No credit card is required for Cloud. A Business trial lasts 14 days with all Business features and does need a credit card.

Not included in the Cloud trial: admin roles, the project Editor role, named versions, workflow diffs and the n8n public API.

Data location and security

Hosted plans store data in the EU, on servers in Frankfurt, Germany. Self-hosted data lives where you put it. Starter and Pro include an encrypted credential store, credential sharing and AI usage controls. Enterprise adds external secret store integration, execution data redaction, audit logging, log streaming and enforced 2FA.

Assistant credits

The n8n Assistant, still in preview, is included in every Cloud plan except Enterprise. Each plan has a monthly allowance: 1,600 credits on Starter, and 4,000 or 9,600 on Pro depending on plan size. Credits are used as you use the Assistant, and how many depends on the size and complexity of what you are building. The balance refreshes monthly and unused credits do not roll over. You can buy extra credits, which are consumed after the included ones run out. On self-hosted n8n the Assistant is available in preview with your own key, so no credit limits apply.

This is separate from the AI models you call inside workflows. Starter and Pro list "AI models without API keys" and gateway credits, while Enterprise does not include gateway credits. For anything else you connect your own model provider, and that provider bills you. See LLM API prices for what those calls cost.

How to stay under your quota

Because the unit is a workflow run, the cheapest changes are structural.

1. Batch inside one run. One execution can loop over many items. A workflow that fires once per row costs a run per row, while one that fires once and loops costs one.
2. Slow down schedules. Moving a five-minute poll to fifteen minutes cuts that workflow's executions by two thirds.
3. Prefer webhooks to polling. A webhook runs when something happens, and a poll runs whether or not anything happened.
4. Filter early. n8n's saved-execution storage is capped by count and size, and reaching the cap only reduces history, but a workflow that exits early still counts as a run.
5. Check the Insights dashboard monthly. It shows production execution counts so you can see the trend before the limit arrives.

These are our suggestions from n8n's own counting rules, not figures from the pricing page.

Moving between plans

You start on a Cloud trial, move to Starter, then Pro. Past Pro, the choice is Business (self-hosted) or Enterprise. If you already run Community, subscribing to Business gives you a license key by email, which you apply to your instance. n8n will not cancel your other subscriptions automatically. If you are on a Starter, Pro, Startup, Enterprise or other Business plan, you must cancel them yourself, and annual subscribers who want to switch to Business should contact sales.

Cloud against self-hosted

Who should pick which

Our take

n8n's pricing is easy to explain and easy to model, which is not true of most automation tools. Counting whole workflow runs makes the monthly bill depend on how often things happen, not how clever the workflow is. A scheduled workflow every five minutes is the classic way to burn a quota, so check schedules before picking a plan.

The big jump is Pro at 50 euros to Business at 667 euros. There is nothing between them, and Business is self-hosted. For a small team that wants sharing and hosted infrastructure, the real choice is Pro with its 10K ceiling or Enterprise by quote. We have not run n8n in production for this page, so we make no claim on performance.

For related options, see the AI agent platform overview and OpenAI Agent Builder. For the model costs your workflows will incur, see LLM API prices.

Checklist

• List every workflow and its trigger, then count monthly runs.
• Check schedules: every 5 minutes is about 8,700 executions a month alone.
• Decide whether you need hosted or can self-host.
• Confirm whether you need sharing, SSO or Git; Community has none.
• Remember that LLM tokens are billed by the model provider, not by n8n.
• Check the monthly price toggle if you do not want to pay annually.

Questions we hear

Can I run n8n for free in production? The Community edition is free, includes almost the complete feature set and runs without a license key. You pay for your own server. It lacks sharing, SSO, Git version control, projects and other features listed above.

What about workflows that fail? n8n's stated principle is that you pay when a workflow runs from start to finish. The page does not define how failed runs are counted.

Is the Pro plan enough for AI agents? Agents, MCP and AI nodes are included from Starter. The limits that bite are execution count, concurrency and the five-minute duration on Starter, which is 40 minutes on Pro.

Can I mix plans? Subscribing to Business does not cancel existing subscriptions; you cancel those yourself.