n8n Pricing 2026: Cloud $24/mo, Self-Hosted License Free

If you searched for the n8n price, you have probably seen three different numbers for the same plan. You are not misreading. n8n's pricing page shows euros to some visitors and dollars to others, the Business plan is a license for your own server rather than a Cloud tier, and what used to be called "Embed" now sits on a page called OEM with no public price.

I set up self-hosted n8n for clients, and I have deployed client stacks on both Oracle Cloud and Hostinger, so the number I care about is the real monthly bill, not the headline. Below are the figures that decide that bill, taken from n8n's pricing page, its docs and its help center, with the currency each one was shown in and what it buys you. Prices verified 2026-09-19.

Want to see n8n doing real work first? Run my Google Maps lead finder or the SEO audit, both live n8n workflows.

n8n pricing at a glance

A few things the table cannot show. The pricing page says every plan includes unlimited users and workflows and every integration, so n8n prices on executions only. Annual billing saves 17%: Starter is $240 a year instead of $288, and Pro is $600 instead of $720. Execution quotas belong to paid plans; n8n's docs say paid plans, cloud or self-hosted, have one, while on the free Community Edition your server is the limit. And tax can be added at checkout depending on your country.

n8n price in USD vs EUR: why Business shows two numbers

Loaded from Doha on 2026-09-19, the pricing page showed euros before its scripts finished loading: 20 EUR, 50 EUR and 667 EUR a month on annual billing. Once the live prices arrived, Starter and Pro switched to $20 and $50. Business showed 667 EUR on one load and $800 on the next two.

n8n's own checkout price data explains the Business gap. The pricing page fetches live prices from n8n's payment provider, and for Business that data sets a euro price with a fixed dollar price for one country only. The 40,000-execution tier is 800 EUR a month, with a $960 override for buyers in the United States. Queried from Qatar the same day, the same checkout converted 800 EUR to $918.32 a month instead. The pricing page can still show you $960 outside the US, so treat that figure as the US price, not yours.

What that means for your budget:

• Starter and Pro carry the same number in every currency. Starter is $24, 24 EUR or 24 GBP a month and Pro is 60 in the same currencies, before tax. Nothing to convert.
• Business in the US is a fixed $960 a month on monthly billing, or $800 a month on annual billing ($9,600 a year).
• Business everywhere else starts from the euro price: 800 EUR a month, or 667 EUR a month on annual billing (8,000 EUR a year). A dollar figure outside the US is a conversion, so it moves with the exchange rate.
• Annual Business can be paid in either currency by invoice. n8n's pricing FAQ offers a self-service DocuSign contract for invoice or wire payment, with one form for euro payment and one for dollar payment.
• Check the currency and the total on the checkout screen before you pay, not on a blog post, including this one.

So when a page quotes Business at "$800 a month", it is either the annual US rate or the monthly euro figure with a dollar sign on it.

What counts as an execution, and what does not

n8n bills on production executions. The pricing page defines an execution as "a single run of your entire workflow", however many steps it has and however much data it processes. The executions docs spell out how each trigger counts:

• Schedule trigger: one execution every time it fires, whatever the outcome.
• Polling trigger: one execution only when it finds new data.
• Webhook trigger: one execution for every inbound request that activates it.

Five kinds of run do not count toward your quota:

• Manual executions
• Sub-workflow executions (only the parent, top-level execution counts)
• Error workflow executions
• Polls that return no data
• Malformed or rejected requests

Two practical consequences. A parent workflow that calls five sub-workflows is billed as one execution, not six. And a schedule that fires every 5 minutes uses 8,640 executions in a 30-day month on its own, more than three times Starter's 2,500. On Cloud, every run past the quota fails, as the next section explains.

n8n subscription cost: Cloud Starter and Pro

These are the two plans n8n hosts for you at a public price.

Starter: $24 a month, or $240 a year. 2,500 executions a month, 5 concurrent executions, 1 shared project, 2,300 AI credits a month and forum support. Cloud keeps up to 2,500 saved executions (2.5 GB), with 7 days of execution logs and 1 day of workflow history.

Pro: $60 a month for 10,000 executions, or $144 a month for 50,000. Annual billing brings those to $50 and $120 a month ($600 and $1,440 a year). You get 20 concurrent executions (50 on the 50K option), 3 shared projects, admin roles, global variables, workflow history, execution search, 7 days of insights, and 5,700 or 13,700 AI credits a month depending on the option. Cloud keeps up to 25,000 saved executions (25 GB), with 30 days of execution logs and 5 days of workflow history.

The 50K Pro option is easy to miss, because the Pro card shows the 10K price until you open its dropdown. If you run between 10,000 and 50,000 executions a month and do not need SSO or Git environments, $144 on Cloud is far below the cheapest Business tier.

The free trial runs for 14 days with most Pro features. n8n's pricing FAQ lists what it leaves out: admin roles, the project Editor role, named versions, workflow diffs and the n8n public API. It is capped at 1,000 executions, 5 concurrent executions, a 180-second execution timeout and 1 team project, on the same computing power as Starter. No credit card is needed, and the trial expires without charging you. The trial docs say n8n deletes the workspace when the trial expires, and also that you have 90 days to download your workflows afterwards. Export anything you want to keep before day 14 rather than rely on that window.

Four smaller points are worth knowing before you pay. Runs beyond your plan's concurrency limit wait in a queue and then run (concurrency docs); that is a different limit from your monthly quota. You cannot buy extra AI credits yet. Saved-execution limits do not stop your workflows, they only change how much history is kept. And hosted plans store data in the EU, on servers in Frankfurt, Germany.

What happens when you run out of executions

On Cloud, the monthly quota is a hard stop, not an extra charge. n8n's Help Center says: "Executions at the limit are not queued. They fail immediately with an error." Your workflows stay active and your triggers keep firing, but each run fails with "Execution limit reached. Consider upgrading your plan." Runs that fail this way stay failed and do not re-run on their own. Quotas reset at the start of each calendar month, not on your billing date.

Before that reset you have three ways out:

• A one-time courtesy reset. n8n offers it once per customer on the trial, Starter and Pro 10K.
• An upgrade. The Help Center says it takes effect immediately.
• Extra executions on monthly Pro 50K (Pro-2 in n8n's Help Center). You can raise the limit yourself in the Admin Panel, in preset tiers up to 500,000 a month. The Help Center does not list prices for those tiers. On annual Pro 50K, support makes the change for you.

Business works differently: workflows keep running and overage may be invoiced, covered in the next section.

n8n license cost: what self-hosting legally costs

Self-hosted n8n is one install. Without a license key it runs as the free Community Edition, and adding a Business or Enterprise key switches on those editions (hosting docs).

Can you get the Business plan on n8n Cloud?

No. n8n's Help Center says "you cannot upgrade an existing Cloud plan to the Business plan" and that Business "is a self-hosted only plan and it is not available on the Cloud at this time." The start-up plan page says the same. You buy a license key and apply it to your own server. The hosted plans are Starter, Pro and Enterprise.

The feature split comes from n8n's Community Edition features page, which also says you can stay on the free edition indefinitely.

One detail if you read the source code: files with ".ee." in their filename, or ".ee" in their folder name, are not covered by the Sustainable Use License. They sit under n8n's enterprise license, which lets you copy and modify that code for development and testing without a subscription, but requires a valid license for your usage before you run it in production.

Business license tiers

The Business card on the pricing page has an execution dropdown. The euro columns are n8n's base price. The dollar columns are the fixed prices n8n's checkout data sets for US buyers.

n8n's pricing FAQ fills in how the license works day to day.

• n8n pitches Business at companies with fewer than 100 employees.
• After payment, the license key arrives by email and you apply it to your existing Community instance. The key must ping n8n's license server daily to stay active. The license docs cover activation in the UI or through the N8N_LICENSE_ACTIVATION_KEY environment variable (which has no effect if the instance already has an activated license), plus N8N_LICENSE_CERT for air-gapped setups. Set one or the other, never both. The license server runs on Cloudflare, so a locked-down firewall has to allow Cloudflare's full IP range.
• One key works on an unlimited number of instances. Usage from all of them counts toward one quota.
• Go over your quota and workflows keep running, but overage charges may apply if you do not move up a tier. Overage is sold in buckets: 4,000 EUR for each extra bucket of 300,000 executions. The FAQ does not say whether a partly used bucket is prorated, so plan as if a small overage could cost a whole bucket. n8n's sales team contacts you as you near the quota, and if there is no response or upgrade, the overage is invoiced 45 days after it occurs. On annual plans, n8n reaches out when you get close to 80% of your annual quota.
• The Business free trial gives you all Business features for 14 days. Unlike the Cloud trial, it needs a credit card.
• Support is the community forum. Dedicated support is Enterprise only.
• Annual Business and Enterprise can be paid by invoice and wire transfer. On an annual commitment, cancelling takes effect at the end of the term, and n8n does not give pro-rata refunds.

Start-up plan

n8n advertises the start-up plan as 50% off Business for companies with fewer than 20 employees and less than 5M EUR in total funding. It is self-hosted, like Business. n8n's checkout data prices the 40,000-execution tier at 400 EUR a month, or 4,000 EUR a year (333 EUR a month). US and Canadian start-ups get a fixed dollar price of $400 a month, or $4,000 a year ($333 a month).

That is exactly half the euro Business price. Against the $960 US Business price, $400 works out to about 58% off. You can keep the discount for a year, or longer while you still qualify.

Enterprise

No public price. It can be hosted by n8n or run on your own servers. n8n invoices on Net15 terms, paid annually in advance, and SLA support contracts are an extra cost.

n8n embed pricing (now called OEM)

Search for n8n Embed today and you land on n8n's OEM page. The "Embedding n8n" item in n8n's site menu points there, and the old n8n.io/embed/ URL returned a 404 on 2026-09-19.

n8n's OEM page and docs set out these terms:

• No public price. Pricing is based on executions: you commit to a yearly volume of executions across all your end customers and are billed against that commitment.
• Executions are pooled across all your end customers under a single contract.
• It is separate from Enterprise, with its own commercial terms, a dedicated sales team and platinum support.
• n8n branding stays visible in the editor. White-labeling is not offered.
• It needs its own contract. The OEM docs say an OEM deployment requires a separate commercial agreement with n8n.
• There is no official minimum. A figure of about $50,000 a year circulates in n8n community forum threads and third-party articles. Neither n8n's OEM page nor its OEM docs publish that or any other number. Get a quote.

Before you ask for a quote, check whether you need OEM at all. n8n's own quick test: if your end users only trigger and consume workflows you built, that is a backend use case. If they create and edit their own workflows inside your product, that is OEM.

For the backend case, n8n's pages do not say the same thing. The license FAQ says backend use is usually allowed "as long as the back-end process doesn't use users' own credentials to access their data." The OEM docs say backend use is "available on all paid plans, under the standard license, with no separate agreement needed." The OEM page says it "runs on a regular Enterprise license." If n8n powers a paid feature of your product, email license@n8n.io, explain whose credentials the workflows use, and get the answer in writing before you ship.

If you are building that product around n8n, the app side is what I do in full-stack AI app development: the front end, the database, and n8n running the background work.

Is n8n free?

The software is. n8n Cloud is not. The self-hosted Community Edition costs $0 to license, with no execution quota from n8n. n8n's pricing page listed no free Cloud plan on 2026-09-19: the cheapest hosted option is Starter at $24 a month, after a 14-day trial that is capped at 1,000 executions and needs no credit card.

Is n8n free for commercial use?

For internal business use, yes, and the Community license costs $0. n8n's Sustainable Use License is a fair-code license, and n8n says plainly that it does not call itself open source. The core rule: you may use or modify the software "only for your own internal business purposes or for non-commercial or personal use."

n8n's license page lists these as allowed:

• Syncing data your company controls, for example from a CRM to an internal database
• Consulting services around n8n, such as building workflows
• Setting up and maintaining n8n on an internal company server
• An embedded AI chatbot that runs on your company's own credentials

It lists these as not allowed without a separate commercial agreement:

• White-labeling n8n and offering it to your customers for money
• Hosting n8n and charging people money to access it
• Collecting your end users' own credentials (their HubSpot login, in n8n's example) so n8n can sync their data into your app

Two of the allowed examples cover my own work: consulting such as building workflows, and setting up or maintaining n8n on a company's own server. n8n's license page adds that you are "free to offer commercial consulting or support services" without a separate agreement. That is why everything I build goes into accounts the client owns. Running one instance myself and charging clients to log in would count as hosting n8n and charging people to access it, which needs a commercial agreement. For anything outside these lines, the contact is license@n8n.io. The full text is in the LICENSE.md file on GitHub.

What self-hosting n8n really costs

The license is $0. You still pay for the server and the hours to run it.

n8n itself is light. Its AWS guide lists the memory n8n Cloud gives the n8n container: 320 MB on Starter, 640 MB on Pro 10K and 1,280 MB on Pro 50K, before your database and operating system. n8n's OEM prerequisites page gives an illustrative range based on n8n Cloud, 320 MB to 2 GB of memory and 512 MB to 4 GB of SSD for the database, and says small instances should be enough for most use cases because n8n is not CPU intensive. The bigger number is in n8n's Docker Compose guide: its full setup, which includes the sandbox that powers n8n Assistant, asks for at least 4 GB of RAM and 2 vCPUs. If you follow that guide, size the server to match.

Here is what that costs on the two platforms I use for client stacks, plus DigitalOcean as a pay-monthly option. Prices in USD:

Sources: Oracle Always Free resources, Hostinger n8n hosting, DigitalOcean droplet pricing.

Oracle is the lowest possible bill, and my step-by-step is in self-hosting n8n on Oracle Cloud's free tier. Hostinger is the simpler paid route, covered in self-hosting n8n on a Hostinger VPS. One of my Oracle builds is written up as a case study: a $0-a-month self-hosted AI stack. It was built in late 2025 on 4 OCPUs and 24 GB, the free allowance at the time. Oracle cut the Always Free Arm allowance to 2 OCPUs and 12 GB in mid-2026, so size a new build to the current limit.

The server price also leaves out four costs:

• Your time. Updates, backups, HTTPS, monitoring and fixing things when they break. My production checklist for self-hosted n8n covers each one.
• Downtime risk. n8n's docs say executions missed while your instance is down or restarting, such as scheduled or webhook runs, are not recoverable.
• Paid features. SSO or Git environments on your own server still need the Business license on top of the server bill.
• AI usage. On self-hosted n8n the AI Assistant is in preview and you bring your own API key, so n8n's credit limits do not apply and your model provider bills you for the usage.

Side by side: Cloud Pro at 10,000 executions is $60 a month, and a Hostinger KVM 2 at its renewal price is $14.99 a month with no execution quota from n8n. The difference is what you pay n8n to run the ops for you.

n8n vs Zapier vs Make on the same workload

The three tools count different things.

• n8n bills per execution: one full workflow run, trigger to finish.
• Zapier bills per task. Zapier's help center counts every successful action as a task. Trigger steps, Filter and Paths steps, actions that error or halt, and steps in Zapier's own utility apps such as Formatter and Delay do not count.
• Make bills per credit. Make's pricing page counts each module action as one credit, reading data from an app or webhook included. The router module costs nothing.

Take a lead workflow with a webhook trigger plus 4 app actions (enrich, score, write to CRM, notify Slack) that runs 2,000 times a month. The Zapier figure assumes none of the four is a Formatter or other free utility step.

This is an illustrative example using prices checked the same day on each vendor's page (Zapier pricing, with Make and n8n linked above). Your real cost depends on your steps, volume and retries.

So n8n is not automatically the cheapest. At this volume Make's Core plan costs less, but 10,000 credits is its whole allowance. One extra module takes you past it, and then you buy extra credits or move up a tier; Make's pricing page says scenarios stop running until credits are added again, while incoming webhooks queue up to your webhook queue storage. n8n Cloud has the same kind of wall: at the quota, executions fail until the month resets or you upgrade. Zapier costs the most here because every successful app action is a task. n8n pulls ahead as workflows get longer, because 4 steps or 40 steps is still one execution.

For a fuller feature comparison, see n8n vs Zapier for business. To model your own volume, my n8n cost calculator does the rough math.

How to lower your n8n bill

• Pay annually if you are staying. It saves 17% on Cloud: Starter drops from $288 to $240 a year, Pro from $720 to $600.
• Price Pro 50K before Business. $144 a month buys 50,000 Cloud executions, far below the cheapest Business tier.
• Use sub-workflows for work that runs off the same trigger. Only the parent execution counts, so one parent calling several sub-workflows is billed once.
• Prefer a polling trigger to a tight schedule where the app offers one. A polling trigger counts only when it finds new data; a schedule counts every time it fires.
• Test with manual runs. Manual executions from the editor do not count toward the quota.
• Check the start-up plan if you have fewer than 20 employees and less than 5M EUR in funding. n8n's pricing FAQ says the start-up page is where you retrieve the discount code.
• Self-host the Community Edition if you can run a server: no license fee and no execution quota from n8n.

Which n8n option fits you

• Under 2,500 runs a month and no appetite for servers: Cloud Starter.
• Up to 50,000 runs and you still want it hosted: Cloud Pro. Price the 50K option at $144 first. On monthly billing it can be raised in preset tiers up to 500,000 a month, at prices n8n's Help Center does not list.
• You need SSO, Git environments or shared projects and can run your own server: the Business license, or the Start-up plan if you qualify.
• You want it hosted with more than Pro offers (200+ concurrent executions on n8n Cloud, log streaming, an SLA, invoice billing): Enterprise, by quote.
• You are comfortable with Docker and want the lowest bill: Community Edition on a VPS or on Oracle's free tier.
• Your customers will build workflows inside your product: OEM. Talk to n8n before you build.

My default: if you are comfortable with Docker, Community Edition on a VPS with 2 vCPUs and 4 GB of RAM. If you are not, Cloud Starter until you outgrow it, then price the 50K Pro option before you look at anything self-hosted.

If you are still working out what n8n is before you price it, start with what n8n automation is and what it is good for.

When to hire help instead of doing it yourself

Most simple automations, you should build yourself. n8n is approachable, and a two or three step workflow is a good first build.

Paying someone starts to make sense in three situations.

1. The workflow touches money, customers or compliance. A lead flow that silently drops records can cost more than the build did. Production workflows need deduplication, retries and error alerts that are easy to skip while you are learning.
2. You are connecting several systems with real edge cases. A CRM, billing, a database and notifications, with branching logic and failure recovery, is where DIY projects tend to stall, and where a self-hosted instance needs proper backups and monitoring.
3. Your time is the bottleneck. If the hours you would spend learning error handling and backups are worth more spent on sales or delivery, hand the build off.

For reference, my fixed-price CRM automation package with n8n starts at $299 on my Upwork catalog, with larger tiers at $849 and $1,999. One of my CRM builds on Supabase and n8n is written up as a case study. Custom work is scoped and priced in writing within 24 hours of a call. More on how I work as an n8n consultant.

Key takeaways

• Cloud Starter is $24 a month and Pro is $60 ($20 and $50 on annual billing), with a $144 Pro option for 50,000 executions.
• Business is a self-hosted license, not a Cloud plan. It starts at 800 EUR a month, or a fixed $960 for US buyers, so the currency matters for Business in a way it does not for Starter and Pro.
• The Community Edition license costs $0, and registering it with an email adds folders, debug in editor and custom execution data for free.
• Cloud quotas are a hard stop. Past the limit, executions fail until the month resets or you upgrade. Business keeps running and may bill overage.
• Embed is now OEM, with no public price. You commit to a yearly execution volume. Backend-only use is a different path, so confirm which one you are with n8n.
• Only production executions count. Manual runs, sub-workflows and error workflows do not.
• Self-hosting costs a server and your time. $0 on Oracle's free tier, $6.49 to $14.99 a month on Hostinger, or $24 a month for a DigitalOcean droplet that meets n8n's Docker Compose guide.

If you want a second opinion on Cloud vs self-hosting for your actual volume, book a 30-minute call and we will size it together.

Sources

• n8n pricing: plan prices in USD and EUR, limits, AI credits, trial, overage, license key and billing FAQ
• n8n start-up plan: eligibility, start-up prices, Business being self-hosted only
• n8n Help Center: running out of executions, adding executions on Pro 50K and changing your Cloud plan
• n8n Sustainable Use License docs and LICENSE.md: commercial-use rules and examples
• n8n enterprise license: terms for ".ee." source files
• Community Edition features and choose how to use n8n: free, registered, Business and Enterprise editions
• Manage your license: license activation options
• Understand executions and concurrency: what counts toward the quota
• Cloud free trial: trial length and limits
• n8n OEM and OEM deployment docs: embed terms and backend use
• AWS sizing, OEM prerequisites (illustrative sizing) and Docker Compose guide: memory, CPU and storage guidance
• Oracle Always Free resources, Hostinger n8n hosting, DigitalOcean droplets: server prices
• Zapier task usage, Zapier pricing, Make pricing: comparison figures

All prices verified 2026-09-19 against the pages above, checked from Doha, Qatar. Pricing changes, so check n8n's page and your checkout currency before you buy.