Fast answer
The n8n workflow cost calculator estimates your true monthly automation cost across the four layers that actually drive it: execution volume, hosting choice, AI model calls, and maintenance time, then compares the total against what the same workflows would cost on Zapier or Make. The methodology reflects how n8n pricing really works. n8n charges per workflow execution rather than per step, which is why multi-step workflows that get expensive on task-priced platforms often run far cheaper on n8n. You enter monthly executions, average steps, and how many workflows call AI models, then pick a hosting path: n8n Cloud starting around 25 dollars per month, or self-hosting on a 10 to 40 dollar VPS plus the admin hours that actually dominate self-hosted cost. The output is a monthly total, a per-execution cost, and a side-by-side against Zapier and Make at your real volume, which is the comparison that matters, since platform rankings flip depending on steps per run.
On this page
What this page covers
A tool visitor should leave with a decision, not just a number: build now, prepare first, choose another workflow, or follow a course path.
- 01Runs
- 02Workflows
- 03AI calls
- 04Hosting
- 05Comparison
- 06Recommendation
Why does this matter now?
The subscription price is the least interesting number in an automation stack, and comparing platforms by it produces bad decisions in both directions. Zapier bills per task, Make per operation, and n8n per execution, so the same workflow can differ in cost by five to ten times across platforms depending on how many steps each run contains. Self-hosting muddies it further: the server is cheap, but updates, backups, and debugging are paid in hours. Agencies quoting client work without this math either underprice and eat the difference or overprice and lose the deal.
Internal path
Where to go next from this page
These links are part of the A8gent learning and conversion path. Use them to move from concept, to diagnosis, to workflow build, to course.
What you should be able to do after this
- Estimate n8n cost
- Compare Zapier and Make
- Budget AI calls
- Choose hosting
How do you do it, step by step?
1. Count executions and steps per workflow
Estimate how many times each workflow runs per month and how many steps a typical run contains, including branches and error retries. This distinction is the whole game: n8n charges per execution regardless of steps, while Zapier and Make charge per step or operation. A 15-step workflow running 1,000 times a month is where the platforms diverge dramatically.
2. Choose the hosting path with honest labor costs
n8n Cloud starts around 25 dollars per month and removes infrastructure work. Self-hosting needs a 10 to 40 dollar VPS plus Docker upkeep, updates, backups, and security patching, realistically two to five admin hours per month. Price those hours at your real rate before concluding self-hosting is cheaper. For one or two workflows, it usually is not.
3. Add AI model costs separately
n8n does not bill for your model usage; your OpenAI or Anthropic account does. Estimate calls per execution, including retries and multi-step agent nodes, times average prompt size, times the per-token price. AI-heavy workflows often spend more on model calls than on n8n itself, and this line is the most common surprise on the first real invoice.
4. Budget maintenance and debugging time
Workflows break when APIs change, credentials expire, and edge cases appear. Plan two to five hours per month across a small workflow portfolio, more if you self-host. For agencies, this is billable retainer scope; for internal teams, it is someone's job that must be named. Unmaintained n8n instances fail quietly until a client notices missing data.
5. Compare against Zapier and Make at your actual volume
Run the same volume math on all three: executions times steps against Zapier's task pricing and Make's operation pricing. n8n typically wins on multi-step, high-volume workflows, while Zapier can win for a handful of simple zaps where its higher per-task cost buys zero infrastructure thought. Decide on your numbers, not on platform marketing.
6. Re-check the math as volume grows
Cost rankings flip with scale. A stack that is cheapest on n8n Cloud at 5,000 executions may justify self-hosting at 50,000, and a client project that fit Zapier's free tier can become its most expensive line item. Re-run the calculator quarterly or whenever a workflow's volume doubles, and before quoting any new client retainer.
What mistakes should you avoid?
- Comparing platforms on subscription price when per-step versus per-execution billing is what actually moves the total.
- Forgetting that AI model calls are billed by the provider on top of n8n, often exceeding the platform cost.
- Calling self-hosting free while ignoring the two to five monthly admin hours it actually consumes.
- Quoting client automation work without a maintenance line, then absorbing every API change unpaid.
- Never revisiting the math as volume grows, missing the point where a different hosting tier or platform becomes cheaper.
FAQ
Is n8n cheaper than Zapier?
Usually yes for multi-step workflows at volume, because n8n charges per execution while Zapier charges per task, so a 15-step workflow costs one n8n execution but 15 Zapier tasks. For a few simple, low-volume zaps, Zapier's convenience can be worth its premium. Run both pricing models at your real monthly volume before deciding; the ranking depends entirely on steps per run.
How much does it cost to self-host n8n?
The server is cheap: a capable VPS runs 10 to 40 dollars per month, and the community edition is free to run. The real cost is labor - Docker updates, backups, SSL, security patches, and debugging - typically two to five hours monthly. Priced honestly, self-hosting beats n8n Cloud mainly at higher volume or when data residency requires it.
How does n8n pricing work?
n8n Cloud bills by workflow executions per month, starting around 25 dollars, and an execution counts once no matter how many steps it contains. That per-execution model is n8n's core cost advantage over per-task and per-operation competitors. Self-hosted community edition has no license fee for standard use; you pay for infrastructure and your own time instead.
Should an agency self-host n8n for clients?
Only if infrastructure is genuinely part of your offer: you can commit to updates, backups, monitoring, security, and support response times, and you price that into retainers. Otherwise use n8n Cloud or a managed host and spend your hours on workflows, which is what clients pay for. A neglected self-hosted instance is a client-losing incident waiting to happen.
What makes n8n workflows expensive to run?
Three things, roughly in order: AI model calls billed by your provider on top of n8n, especially agent nodes with retries and large prompts; maintenance hours when APIs and credentials change; and execution volume growth pushing you into higher plan tiers. The platform fee itself is rarely the problem, which is why calculating all four layers matters.
Sources & further reading
Was this page helpful?