The same 20 workflows cost $16 a month on Make, about $55 on n8n and $129 on Zapier. Identical automations, three meters, an eight-fold spread.
n8n, Zapier and Make meter that same stack differently, and that single fact decides more of this comparison than any feature list. Zapier counts tasks, one for every billable action a Zap completes, while triggers and filters run free. Make counts credits, charging for every module that executes, trigger included. n8n counts executions, one charge for a complete run of the whole workflow regardless of step count.
Price one real 20-workflow stack on all three meters and the ranking is not the one most teams assume going in. Make comes out cheapest, n8n second at about $55, and Zapier roughly eight times the price of Make for running the identical workflows. That is not a universal law, it falls out of one specific mix of steps and volume, and the ranking moves once that mix changes.
We reuse the same 20-workflow stack from the Zapier comparison so all three meters are priced on identical work, from an agency running 200+ n8n workflows in production and building on all three platforms. This piece prices that stack on all three meters at once, then gives the decision framework for choosing before a single workflow gets built.
n8n vs Zapier vs Make: what is the actual difference?
Each platform bills a different unit of work, and that unit decides almost every price surprise in this comparison.
- Zapier bills tasks, one for every successful billable action a Zap completes, while triggers, filters, and built-in tools such as Formatter and Paths (Zapier’s built-in data tools) run free.
- Make bills credits, charging for every module that executes on the canvas, the trigger included.
- n8n bills executions, one charge for a complete run of the whole workflow, at any number of steps or any amount of code inside it.
That difference is not a rounding footnote. A twelve-module workflow that Zapier might price at two billable tasks, because most of those steps are free triggers and filters, is priced by Make at roughly twelve credits per run, because every module on its canvas carries a meter. n8n prices the same workflow at one execution, regardless of its step count. Three vendors give three different answers to what one run of the identical automation actually costs.
None of the three units convert cleanly into another. A “10,000” on a Zapier plan label, a “10,000” on an n8n plan label, and 10,000 Make credits are three different amounts of real work. Comparing the plan labels instead of your own workload is the single most common mistake in a three-way evaluation.
What does the same stack cost on all three platforms?
Make is the cheapest at $16 a month, n8n Cloud second at about $55, and Zapier the most expensive at $129, for the identical 20 workflows. Take that stack: 10 lead-intake workflows with 3 billable actions each, running 200 times a month; 6 billing-sync workflows with 2 actions each, running 250 times a month; and 4 reporting workflows with 5 actions each, running 50 times a month. Total volume across the stack is 3,700 runs a month.
On Zapier, triggers are free and only billable actions count, so this stack consumes 10,000 tasks a month: 6,000 from lead intake, 3,000 from billing sync, 1,000 from reporting. That lands on the Professional plan at $129 a month on annual billing.
On n8n, one run of a workflow is one execution no matter how many steps it has, so the same stack is 3,700 executions a month. Starter’s 2,500-execution ceiling will not cover that, so the plan that fits is Pro, at EUR 50 a month on annual billing, roughly $55.
On Make, a credit is charged per module per run, and the trigger counts as a module alongside every action. Lead intake runs 4 modules, 3 actions plus the trigger, 200 times across 10 workflows, for 8,000 credits. Billing sync runs 3 modules 250 times across 6 workflows, for 4,500 credits. Reporting runs 6 modules 50 times across 4 workflows, for 1,200 credits.
That totals 13,700 credits a month, past the 10,000-credit rung and into the 20,000-credit rung at $16 a month on annual billing. That total assumes one record per run. A Make module fed a batch of 20 records bills 20 credits, and routers and error handlers are free, so a stack that fans out over lists moves this number fast.
All three prices are as published in August 2026 against each vendor’s own pricing page.
When does that ranking flip?
Three mechanisms flip the ranking between n8n, Zapier and Make, each a matter of scenario shape rather than opinion. Long workflows punish Make first, because every extra module on the canvas multiplies the credit cost of every run. A workflow with twenty modules costs five times what a four-module workflow costs at the same volume, while n8n’s per-run execution charge does not move as step count grows.
Very short workflows at very high volume punish n8n Cloud instead. Pro’s execution ceiling stops at 10,000 executions a month, and the next tier, Business, jumps to EUR 667 a month, more than thirteen times the Pro rate. A three-step workflow running tens of thousands of times a month can hit that ceiling long before Make’s credit meter starts to hurt.
AI steps punish Zapier specifically. Its June 2026 pricing change bills the default Advanced tier at 3x, and each tool call at that same 3x, so one agent step with four tool calls comes to 3 for the base plus 12 for the tool calls: fifteen tasks for a single step. n8n meters that same agent workflow at one execution regardless of how many tools it calls in a single run.
Triple the same 20-workflow stack and the ranking does not just compress, it flips.
| Volume | Make | n8n Cloud | Zapier |
|---|---|---|---|
| 1x (3,700 runs a month) | $16 (13,700 credits) | about $55 (3,700 executions) | $129 (10,000 tasks) |
| 3x (11,100 runs a month) | $55 (41,100 credits) | about $734 (11,100 executions) | between $129 and $289 (30,000 tasks) |
n8n’s 11,100 executions exceed Pro’s 10,000-execution ceiling, forcing the jump to Business at EUR 667 a month, about $734: n8n goes from second-cheapest to the most expensive platform by far. Make’s 41,100 credits cross into the 80,000-credit rung at $55, still the cheapest of the three even at triple volume. Zapier’s 30,000 tasks sit between the $129 rung at 10,000 tasks and the $289 rung at 50,000, so the true number lands somewhere on that curve rather than at either end. Self-hosting is what removes n8n’s cliff: its fair-code license lets a team run n8n on its own server, trading the execution meter for the cost of the box and whoever maintains it.
Some flips are not about price at all. Zapier still wins on connector breadth and its free 100-task tier, worth more to a team testing an idea than the dollar difference at low volume. n8n includes unlimited users on every paid tier, where Zapier prices multi-user editing into a separate Team plan once a second person needs access. Self-hosting n8n removes its execution meter entirely, trading it for real ops work: someone has to patch the server, back it up, and notice when it stops.
Make vs Zapier: which one if n8n is off the table?
Make is usually cheaper at mid volume on short scenarios; Zapier wins on connector breadth, its free 100-task tier and not billing failed or filtered runs. The two meters still differ from each other more than either differs from n8n, so ruling out n8n does not make the choice easy. Scenario shape decides which one wins, not brand reputation.
Zapier gives triggers, filters and formatters away free, so the defensive scaffolding around a fragile integration costs nothing to add: an API outage that burns 400 retries bills zero tasks. Make charges a credit for every module that executes, the trigger included, so a long scenario multiplies its bill with every step someone adds, which is exactly the shape of workflow that costs least on Zapier. Neither meter is wrong, they just reward opposite scenario shapes.
The gap is concrete at small scale. A scenario with three billable actions running 500 times a month is 1,500 tasks on Zapier, which pushes past the 750-task entry rung to the 2,000-task rung at $49 a month on annual billing. The same scenario on Make is 2,000 credits (three actions plus the trigger, times 500 runs), comfortably inside the entry 10,000-credit rung at $9. Five times the price for identical work, before a single AI step enters the picture.
Connector breadth is the other factor, and it belongs to Zapier outright. Zapier’s library is the deepest of the three, which matters most the day a workflow needs an app Make does not support. The free 100-task tier adds to that: room to test an idea before paying for either platform.
Which platform should a non-technical team pick?
A non-technical team should start on Make or Zapier, not n8n, regardless of which platform prices cheaper on a given stack. Both platforms are built to be read by the person who maintains the automation: a visual canvas that shows what a scenario does at a glance, a wide connector library, and nothing to host or patch. n8n rewards a team with an engineer on staff, or an agency running the build. Its real advantages, the code node, where custom logic is a few lines of JavaScript, self-hosting, and AI agent nodes that make tool-calling decisions, all assume someone comfortable debugging a workflow that broke overnight.
The honest reason to weight maintainability over price is durability, not preference. A workflow that only one absent contractor can read is a liability whatever it costs to run. That risk shows up months after launch, when the person who built it has moved on and nobody else can open the canvas.
Hiring an agency to run n8n changes this calculation, because the agency becomes the maintainer instead of one in-house person. That is why teams often start on Make or Zapier and move to n8n only once an engineering relationship, in-house or agency, already exists.
How do you decide in ten minutes?
Three steps turn this from an opinion into a number specific to your own stack.
- Count the billable actions in your five busiest workflows, ignoring triggers and filters, then multiply each by how many times it runs a month. A lead-intake workflow might run 3 billable actions 200 times a month; a billing-sync workflow might run 2 actions 250 times a month; a reporting workflow might run 5 actions 50 times a month. That count is Zapier’s unit; Make bills the trigger as a module too, and the comparator adds it for you.
- Price that total on all three meters at once with the cost comparator, which converts the same run count into Zapier tasks, Make credits and n8n executions so you are comparing real cost, not plan labels.
- Re-run the same count for the stack you expect to be running in twelve months, not the one you have today. A ranking between three billing meters holds only at one exact volume and step count, and shifts as either grows.
Once you know which pair is close, read the deep dive for that pair. n8n vs Make works through the billing model and the crossover point between the two platforms in detail. n8n vs Zapier prices a full year of a 20-Zap stack, including the effect of Zapier’s June 2026 AI task multipliers.
If the ten-minute count says your stack belongs on n8n and you are currently running it through Zapier, migrating off Zapier to n8n is the service that runs the cutover without losing the workflows nobody documented. Every three-way evaluation ends the same way: count your own actions and runs before picking a platform on reputation, because the label on a plan is never the number your workflow actually consumes.
Pricing verified against n8n, Zapier and Make published documentation, August 2026. Both linked articles carry the ladders for their own pair; the Make credit ladder is on Make’s pricing page linked above; all three vendors change pricing without much notice, so confirm current figures before budgeting.