n8n Cloud Starter costs 20 a month billed annually for 2,500 workflow executions, and Pro costs 50 for 10,000, both with unlimited users and unlimited active workflows according to n8n's pricing page. For a five-person business, Starter fits if your workflows run fewer than about 2,000 times a month; otherwise Pro is the safer step.
The Complete Excel Course with AI Integration
From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.
By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner
Buy nowThe price is not per seat, so team size barely matters. What matters is how often your workflows run, because an execution is one run of a whole workflow however many steps it has. At the time of writing (November 2026) the page shows these numbers in euros, and every price in this tutorial is in that currency, so confirm the currency you are quoted at checkout.
What do n8n Cloud Starter and Pro cost, and what do they include?
n8n Cloud Starter costs 20 euros a month billed annually for 2,500 executions, and Pro costs 50 euros for 10,000, both with unlimited users and unlimited active workflows. Business is self-hosted only.
| Starter | Pro | Business | |
|---|---|---|---|
| Price a month, billed annually | 20 | 50 | 667 |
| Executions a month | 2,500 | 10,000 | 40,000 |
| Active workflows | Unlimited | Unlimited | Not listed |
| Users | Unlimited | Unlimited | Unlimited |
| Concurrent executions | 5 | 20 (50 also shown) | Not listed |
| Shared projects | 1 | 3 | 6 |
| Workflow history | 1 day | 5 days | Not listed |
| Overage | Capped | Capped | Not listed |
Three things to note. Business is listed as self-hosted only, with a waitlist for the cloud version, so for a small business the cloud choice is Starter or Pro. The page says annual billing saves 17 per cent, and it shows only the annual prices in the plan table, so read the monthly-billed price off the page's billing toggle if you plan to pay monthly. And the page notes that companies under 20 employees may qualify for a start-up plan with 50 per cent off Business, which is a discount on the self-hosted tier and does not change Starter or Pro.
Pricing is based on workflow executions regardless of complexity. That is the generous part: a workflow with thirty steps and an AI call costs the same one execution as a workflow with two. If you are weighing n8n against other automation tools that charge per step or per task, see Zapier vs Make vs n8n for AI automation, because the counting rules differ.
How do you count your n8n executions before you choose a plan?
Count n8n executions by listing each workflow, estimating how often its trigger fires in a month and adding the totals plus 30 per cent for growth. Choosing a plan comes down to one number, your monthly executions, and you can get a decent estimate before you build anything.
- List every workflow you plan to build in the first six months.
- For each, name its trigger: a form submission, an incoming email, a new row, a schedule, a webhook from another tool.
- Estimate how many times the trigger fires a month. Use real figures: last month's enquiries, invoices, orders.
- Add them up and add 30 per cent for growth.
- Compare with 2,500 and 10,000. Stay under about 80 per cent of the limit.
Schedule triggers need a special warning. A workflow set to check something every 15 minutes runs 96 times a day, which is 2,880 times in a 30-day month, more than the whole Starter allowance for one workflow doing almost nothing. Run it hourly and it is 720; run it twice a day and it is 60. Choose the slowest schedule that still serves the job.
A five-person care agency, counted
An illustration with invented volumes. A five-person home-care agency wants to automate five jobs. Everyone on the team will build or edit workflows, which costs nothing extra because users are unlimited.
| Workflow | Trigger | Runs a month |
|---|---|---|
| New enquiry form to spreadsheet and a reply email | Form submission | 60 |
| Morning rota check and message to managers | Schedule, daily | 30 |
| Supplier invoice PDF read and logged | Incoming email | 120 |
| Visit-note summary for supervisors | New row | 400 |
| Missed-visit alert | Schedule, hourly (8am to 8pm) | 390 |
| Total | 1,000 |
One thousand runs, with 30 per cent headroom, is 1,300: comfortably inside Starter's 2,500. Cost: 20 a month, 240 a year. The same agency on Pro would pay 600 a year for capacity it does not use.
Now grow the agency. It adds a care-plan review workflow that runs per client (180 a month), a hiring-enquiry flow (50) and moves the missed-visit alert to every 15 minutes during the day (48 runs a day, about 1,440 a month). The total becomes 1,000 - 390 + 1,440 + 180 + 50 = 2,280. Add 30 per cent headroom and it is about 2,960, above Starter's 2,500. At that point the cost of Pro, 30 more a month, buys 7,500 extra executions. Per thousand executions, Starter is 8 and Pro 5, so Pro is cheaper per run once you use it.
What else differs between n8n Starter and Pro besides executions?
Beyond executions, n8n Starter and Pro differ in concurrent executions (5 against 20), shared projects (1 against 3) and workflow history (1 day against 5 days), according to n8n's pricing page. Run limits are only part of the comparison, and the page lists other differences that matter to a small team:
- Concurrent executions. Starter lists 5 at once. If a batch of ten form submissions arrives together, some wait. Pro lists 20 (the page also shows 50), which suits busier flows.
- Shared projects. Starter has 1 and Pro has 3. Projects keep workflows and credentials separate, which suits a business with a client-facing set and an internal set. With one project, everyone sees everything.
- Workflow history. The page lists 1 day on Starter and 5 days on Pro. It refers to how long execution history is kept; n8n's data-management documentation gives saved-execution and log-retention figures that differ from the pricing table, so read the current documentation before you rely on history for debugging or audit.
- Active workflows. Both list unlimited. Older write-ups elsewhere mention caps on active workflows, so check the current page before you buy.
If a second project or longer history would save you a real problem, that is a reason to pick Pro even with low volume. If not, volume is the only reason to pay for it.
Four more small businesses, four different answers
- A music school with two admin staff. Enrolment forms (80 a month), weekly reminder emails to parents (40), invoice reminders (60). About 180 runs. Starter, with room to spare, and room to add a feedback survey flow.
- A veterinary practice with three branches. Daily stock-level checks per branch (90), appointment reminders sent by a flow (900), supplier order alerts (40). About 1,030. Starter for now; watch reminders as the client list grows.
- An optician with online booking. Each booking triggers a confirmation, a reminder and a follow-up as three separate runs unless they are built as one workflow with waits. 300 bookings a month times three is 900. Building it as one flow with delay steps may make it 300 executions; check how n8n counts a run that waits before you rely on that.
- A consultancy automating proposals with AI. 25 proposals a month, each triggering a draft, a review request and a send, 75 runs. Tiny. The cost question here is the AI model's usage fees, not n8n.
The third example shows the most useful habit. Because a whole run counts as one execution, a workflow that does many things in one run is cheaper than three workflows that each do one. For invoices as an example, n8n invoice processing: read supplier PDFs with AI shows a single workflow handling the read, the check and the log.
When does n8n Pro become cheaper than Starter?
Pro is cheaper per run than Starter only once you use its capacity: Pro works out at 5 euros per thousand executions against Starter's 8. Take the annual prices at face value. Starter is 20 a month for 2,500 executions. Pro is 50 for 10,000. That only helps you if you use the capacity. A business that runs 1,200 executions a month pays 20 on Starter, or 16.67 per thousand; on Pro the same 1,200 costs 50, or 41.67 per thousand.
The break-even is simple. Pro is worth its extra 30 a month when either of two things is true. Your forecast, with headroom, passes about 2,000 executions a month and Starter would leave you close to the cap. Or you need something only Pro lists, such as the third shared project, higher concurrency or longer workflow history. Between 2,000 and 2,500 runs a month is a grey zone: Starter works until the month a campaign or a busy season pushes you over, which is the month you least want an interruption.
| Forecast runs a month, with headroom | Sensible plan | Annual cost |
|---|---|---|
| Under 1,500 | Starter | 240 |
| 1,500 to 2,000 | Starter, with a monthly check | 240 |
| 2,000 to 2,500 | Pro, unless the volume is steady | 600 |
| 2,500 to 8,000 | Pro | 600 |
| Above 8,000 | Pro with a watch, then a rethink of schedules or tools | 600 and up |
How can you redesign an n8n workflow to use fewer executions?
Redesign an n8n workflow to use fewer executions by running schedules less often and switching to triggers that fire only when something happens. Changing how a workflow is built can move you down a plan, or stop you moving up one. Here is an illustration with invented numbers.
A veterinary practice built a workflow that checks its shared inbox every five minutes for supplier invoices, and another that checks a stock sheet every ten minutes for low items. The inbox flow runs 288 times a day and the stock flow 144, so 432 runs a day and 12,960 in a 30-day month. That is above Pro's 10,000 allowance for two flows that mostly find nothing.
The redesign takes three steps. First, the inbox check moves to hourly during opening hours (12 a day, 360 a month) because invoices are not urgent. Second, the stock check moves to a single morning run (30 a month) because the manager reads it once a day anyway. Third, where the app supports it, the practice switches to a trigger that fires only when something happens, such as a new row or new email, so the workflow runs when there is work instead of when the clock says so. The first two steps alone bring the total to 390 runs a month, a drop of 97 per cent, and the practice fits on Starter with room for more workflows.
The lesson is that executions are a design choice. The question to ask of every workflow is not "can n8n do this?" but "how often does this really need to run?"
How should you check n8n usage in the first month?
Check n8n usage weekly in the first month by recording each workflow's expected runs, comparing them with actual executions and deciding in week four whether to stay, change schedules or move up.
- Week 1. Build one workflow at a time and record its trigger and expected monthly runs in a simple sheet.
- Week 2. Look at the actual executions in n8n and compare with your estimate. If a flow is running twice as often as expected, find out why before adding more.
- Week 3. Total the runs so far and project to month end. Compare with 80 per cent of your plan's limit.
- Week 4. Decide: stay, change schedules, or move up. Write the decision and the reasons in the sheet.
A filled-in version of the sheet, with illustrative figures:
Workflow Trigger Estimate Actual (wk 2 x 2)
Enquiry form to sheet Form 60 52
Supplier invoice reader Email 120 144
Missed-visit alert Hourly 8-8 390 390
Visit-note summary New row 400 338
-----------------------------------------------------------------
Projected month 970 924 (37% of Starter)
The invoice flow ran 20 per cent over estimate because suppliers also sent statements and reminders that the workflow picked up. That is a real, common cause of drift: triggers fire on things you did not think of. A filter step at the start of the workflow does not reduce the executions, since the run still happens, so the cure is a narrower trigger, such as a rule on the mailbox that only forwards invoice emails to the address n8n watches.
If the tracking sheet is the part you keep putting off, try The Complete Excel Course with AI Integration on Coding Liquids.
What costs sit on top of the n8n price?
AI model usage, other apps' own plans, build time and maintenance sit on top of the n8n price. The plan price is not the cost of an automation, and several of these are billed elsewhere or paid in time, so a business owner should keep them in the same budget.
- AI model usage. If a workflow sends text to an AI model, the model's own provider bills for it, separately from n8n. Check the model's current price list and estimate documents a month, as in n8n data entry automation, where each form or email may pass through a model.
- Other apps. Connecting to a spreadsheet, a CRM, an email service or a messaging app may need that app's paid plan or an API allowance. Read the other tool's limits before you assume the n8n side is the only cost.
- Build time. A first workflow often takes a few hours for a non-technical person, and a complex one more. Value the hours, because they are the largest cost in the first month.
- Maintenance. Apps change and a workflow occasionally needs repairs. Name one person who owns each workflow.
- Credentials. Workflows hold access to your tools. Decide who may create or edit them; with unlimited users on every plan, it is easy to give everyone access, and it is not always wise.
With those in the sum, a five-person business that automates five jobs might spend 240 a year on n8n and a similar amount on model usage and extra app plans, plus its own time. The subscription is the small part, and it is still worth getting right.
A sample decision, written down
Whatever you choose, write the decision in a few lines. Six months later, when someone asks why the business is on this plan, the note saves a debate.
Date: [date]
Plan: n8n Cloud Starter, annual
Reason: forecast 1,300 runs a month with headroom; limit 2,500
Re-check: 15th of every month; upgrade if above 2,000 two months running
Owner: [name]
Open question: what happens at the cap (asked n8n support on [date])
If you want help choosing and building your first workflows, I offer one-to-one AI consultations for small businesses.
Should a small business use n8n Cloud or self-host?
A small business without technical staff should usually choose n8n Cloud over self-hosting, because self-hosting costs set-up, updates, backups and security work. n8n can also be run on your own server, and n8n's documentation notes that self-hosting needs technical expertise. Cloud Starter at 240 a year is usually cheaper than the first day a contractor spends on a server. The full comparison is in n8n self-hosted vs cloud costs for a small business, and the Business plan's self-hosted-only status is a sign of where n8n sees larger teams going.
What happens when you hit the n8n execution limit?
At the monthly limit, n8n does not queue new executions: they fail immediately with the message "Execution limit reached. Consider upgrading your plan." The pricing page lists overage on Starter and Pro as capped but does not spell out what that means, and n8n's support article fills the gap: workflows stay active and triggers keep firing, failed executions do not count against the quota, and the count resets at the start of each calendar month whatever your billing date. Starter and Pro customers can ask for one free reset by emailing n8n's help address. Runs that fail are not re-run for you when you upgrade, so a missed run has to be replayed by hand.
If a missed-visit alert or an invoice flow stops silently at 2,500 runs on the 27th, the cost is worse than the plan difference. Protect yourself in three ways. Keep your usage under 80 per cent. Set a calendar reminder for the 15th of each month to check your execution count. And ask n8n support, in writing, what happens at the cap on your plan.
Which mistakes push a small business onto the wrong n8n plan?
A small business ends up on the wrong n8n plan by counting steps instead of runs, setting schedules too often, paying for Pro because of team size or buying annual billing before seeing volumes.
- Counting steps, not runs. A reader sees "fifteen nodes" and imagines fifteen executions. It is one per run.
- Forgetting the test runs. Building a workflow involves dozens of trial runs. n8n's support article says manual runs started from the editor on a paid plan are not affected by the monthly limit, so trial runs while building do not eat the quota; still check your execution count after a heavy build week.
- Setting a 5-minute schedule for a daily job. 8,640 runs a month on one workflow, which would use nearly all of Pro.
- Paying for Pro for the sake of users. Users are unlimited on Starter. Team size is not a reason to upgrade.
- Buying monthly when sure. Annual billing saves 17 per cent, but only commit once a test month has shown real volumes.
The introduction to n8n for small businesses covers whether to use it at all, and the AI guide lists the other tools in this series. Count first, then pick the plan.
n8n Cloud pricing questions
What counts as one n8n execution?
One execution is a single run of your entire workflow, however many steps it has or how much data it handles, according to n8n's pricing page. A twelve-step workflow that runs once is one execution, and the same workflow run by a schedule 96 times a day is 96.
Does n8n charge per user on Cloud?
No. The pricing page lists unlimited users on Starter, Pro, Business and Enterprise, so a five-person team pays the plan price, not five seats. What limits you is executions, concurrent runs and shared projects.
Which n8n Cloud plan fits a five-person business?
Starter, if your workflows run fewer than about 2,000 times a month; otherwise Pro is the safer step. Users are unlimited on both, so team size barely matters and the run count decides. In the illustrative five-person care agency example, about 1,000 runs fit Starter, but growth to roughly 2,960 would need Pro.
Is the Business plan available on n8n Cloud?
Not yet, according to the pricing page: Business is marked self-hosted only, with a waitlist for the cloud version. The page also mentions a start-up discount of 50 per cent off Business for companies under 20 employees, so it does not lower the Starter or Pro price.
What happens when I reach my execution limit?
Executions at the limit are not queued: they fail immediately with an Execution limit reached error, according to n8n's support article. Workflows stay active, and the count resets at the start of each calendar month; an upgrade raises the limit at once. The pricing page lists overage on Starter and Pro as capped, so ask n8n support what applies to your plan.
Further reads
- n8n Cloud or Self-Hosted for Accountants? Costs and Data — Hosting choices for document-heavy firms.
- AI Integration Software for Business: Zapier vs Make vs n8n Costs — Costs of connecting apps across Zapier, Make and n8n.
- n8n or Power Automate? A Microsoft 365 Firm's Cost and Connector Check — If you run on Microsoft 365.
Browse all AI-for-business tutorials
Sources: n8n.io/pricing, which showed euro prices, and n8n's support article on execution limits, checked in November 2026.
