Skip to content

n8n vs Make vs Zapier: Which Automation Tool Wins in 2026

Zapier for non-technical owners, Make for branching logic in the cloud, n8n for volume, data control and agents. What each one bills, and one workflow priced on all three.

In the n8n vs Make vs Zapier choice, pick Zapier when non-technical people will build and own short workflows and you want the most apps with nothing to maintain. Pick Make when your processes branch, loop over lists and reshape data, and you want that on a visual canvas at a lower price than Zapier. Pick n8n when volume is high, data has to stay on your own servers, or the workflow needs code and AI agents, and someone technical will own it.

This page compares all three on what actually decides the bill and the build: what each one counts as a billable unit, what one real workflow costs on each, where the complexity ceiling sits, and what happens when you hit a limit. It also answers Make vs Zapier on its own, and links to the head-to-head pages when you are down to two. Prices are from each vendor's own pricing page, checked on 2 November 2026.

n8n vs Make vs Zapier at a glance

ZapierMaken8n
HostingCloud onlyCloud onlyn8n Cloud, or self-hosted on your own server
BuilderLinear, step by stepVisual canvas with routers and iteratorsNode canvas, code nodes anywhere
Billing unitTask: each successful action stepCredit: each module operation, usually 1 creditExecution: one full run of a workflow, any number of steps
Not billedTriggers, Filter, Paths, Formatter, Delay, Looping, failed actionsRouter and error handler modulesSteps inside a run
Free plan100 tasks a month, two-step Zaps1,000 credits a monthCommunity edition, self-hosted (you pay the server)
First paid planProfessional, $19.99 a month for 750 tasks (annual)Core (price on Make's page, by region)Starter, $20 a month for 2,500 executions (annual)
Apps (vendor's count)9,000+3,000+2,000+, plus any API through the HTTP Request node
Custom codeCode steps with a runtime allowanceLimited, mostly functions and HTTP callsJavaScript and Python nodes
Learning curveLowestModerateSteepest
Best forNon-technical owners, short workflowsOps teams with branching logic, cloud is fineHigh volume, data control, code and agents

Sources: Zapier pricing, Make pricing, n8n pricing and the n8n integrations directory, all checked in November 2026. Make's pricing page loads its plan prices by region, and they did not display when we checked, so this page gives Make's units and rules but not its dollar prices. Check Make's page from your own location before you compare.

You will see very different numbers on other comparison pages: n8n at "400+" or "1,000+" integrations, n8n Cloud at $20, $22 or $24, Zapier at 6,000 or 7,000 apps. Most of those are old snapshots. The figures above are the vendors' own, as of this month, and they will move again.

How each one counts what you pay for

The billing unit matters more than the sticker price, because it decides how the bill grows as a workflow gets longer and runs more often. Many comparisons say every Zapier step costs a task. That is not what Zapier's own pricing page says, and the difference changes the math.

Zapier: tasks

A task is counted each time Zapier successfully completes an action step, including AI by Zapier and Code by Zapier steps. Triggers and polling do not count. Neither do Zapier's built-in tools: Formatter, Paths, Filter, Delay, Looping, Sub-Zap, Digest, Storage, Tables and Forms. Failed actions are not counted either. The Professional plan starts at $19.99 a month billed annually ($29.99 monthly) for 750 tasks; Team starts at $69 a month billed annually for 2,000 tasks. Filters and Paths need Professional or higher.

Make: credits

Make replaced operations with credits as its billing unit. Per Make's help center, one operation equals one credit for non-AI apps, and some AI modules use more. Every module that reads, searches, creates, updates or transforms data uses credits, and that includes the trigger reading from an app or a webhook. The Router module and the error handler modules (Rollback, Break, Resume, Commit, Ignore) use none. The free plan includes 1,000 credits a month.

n8n: executions

On n8n Cloud, an execution is one run of the whole workflow, however many steps it has and however much data it moves. Starter is $20 a month for 2,500 executions and Pro is $50 a month for 10,000, both billed annually (checked November 2026). Business, at $800 a month for 40,000 executions, is a self-hosted plan with the features larger teams ask for. The Community edition, which n8n's pricing page describes as a standard self-hosted version on GitHub, has no execution fee at all: you pay for the server and the time to keep it running. The full plan breakdown is in our n8n pricing guide.

A worked example: one lead workflow, priced three ways

Illustrative example. A website form sends a new lead. The workflow drops test submissions, cleans the phone number, looks the contact up in HubSpot, creates or updates it, routes by region, posts the lead to the rep's Slack channel and creates a follow-up task. Here is what each platform bills for one run.

StepZapier tasksMake creditsn8n
Trigger: new form submission (webhook)011 execution for the whole run
Filter out test submissions0 (Filter)0 (filter between modules)
Clean the phone number0 (Formatter)0 (function in the mapping)
Find contact in HubSpot11
Create or update contact11
Route by region0 (Paths)0 (Router)
Post to the rep's Slack channel11
Create follow-up task11
Per run4 tasks5 credits1 execution

Now the monthly bill at two volumes, from the official price lists (annual billing):

Runs a monthZapierMaken8n Cloud
1,000 leads4,000 tasks: Professional 5,000-task tier, $89 a month5,000 credits: above the free 1,000, so a paid plan1,000 executions: Starter, $20 a month
10,000 leads40,000 tasks: Professional 50,000-task tier, $289 a month50,000 credits: a paid plan sized at 50,000 credits10,000 executions: Pro, $50 a month, with no headroom left

Three things this shows that a list of starting prices hides:

  • Unit counts are not prices. Make bills more units than Zapier for this workflow (5 credits against 4 tasks), yet it usually costs less, because each credit is cheaper. Compare the bill.
  • Zapier's free steps matter. If filters, formatting and routing were billed, this run would cost 7 tasks instead of 4. Design your Zaps around the tools that do not bill.
  • Execution pricing flattens step count and leaves volume alone. Adding five more steps changes nothing on n8n. Doubling the leads doubles the executions. At 10,000 runs, Pro is full: retries and test runs push you past the cap, so plan for the next tier or self-hosting before you get there.

n8n vs Zapier

This is the most common pairing, usually asked by a team whose Zapier bill has started to climb. Zapier wins on time to the first working automation and on app coverage: 9,000+ apps, no servers, and a person with no technical background can own it. n8n wins on cost at volume, on control (self-hosting, code, sub-workflows, error workflows), and on AI agents that call tools.

The real question is ownership. A Zap that a sales manager can fix on a Friday afternoon has a value that does not show on the pricing page. An n8n workflow nobody on the team can read is a liability, however cheap its executions. The full side-by-side, with the switching signals, is in n8n vs Zapier: the complete comparison.

n8n vs Make

Make and n8n are closer than either is to Zapier. Both draw the workflow as a graph, both handle branches, loops and arrays well, and both let you call any API. Make keeps you in a managed cloud and in a visual, mostly no-code way of working. n8n adds self-hosting, real JavaScript and Python nodes, and per-run instead of per-module billing.

Make is the better fit when an operations team builds and maintains its own scenarios and cloud hosting is fine. n8n is the better fit when scenarios loop over hundreds of items per run (each iteration costs credits on Make, nothing extra on n8n), when data must stay in your infrastructure, or when developers will own the workflows. Details in n8n vs Make compared.

Make vs Zapier

If you have already ruled out self-hosting, the decision is Make vs Zapier. Make (called Integromat until its 2022 rebrand) and Zapier are both fully managed cloud services, so the difference is how you build and how you pay.

MakeZapier
BuildingFlowchart canvas: routers, iterators, aggregators on one screenForm-based editor, one step under the other, Paths for branches
Data handlingStrong: arrays, line items, JSON, text and date functions in the mappingFormatter and Looping cover common cases; complex reshaping gets awkward
Apps3,000+9,000+
BillingCredits: trigger and every module, except routers and error handlersTasks: successful actions only; triggers and built-in tools free
Error handlingError routes per module (Rollback, Break, Resume, Commit, Ignore)Failure alerts and replays; less control over what happens next
Who can own itSomeone comfortable thinking in flowchartsAlmost anyone

Choose Zapier when the people building are non-technical, you depend on a less common SaaS app, or the workflows are short (one trigger, two to four actions) and run at modest volume. Choose Make when processes branch, loop over lists or carry many steps per run, and you want the lower running cost that usually comes with it. Make's canvas is also easier to hand to a colleague: they see the whole scenario at once instead of scrolling a list of steps.

One trap on each side. On Zapier, splitting one process across several Zaps to work around the linear editor makes failures harder to trace. On Make, an iterator over a 300-line order costs credits for every line it touches, so a scenario that looks cheap in testing can burn a month of credits on a few large batches.

Where Gumloop fits

Gumloop comes up as the fourth option because it starts from AI agents rather than app-to-app automation: it describes itself as a "multiplayer AI agent builder" where anyone at a company builds agents while IT controls access. It suits teams whose work is mostly reading, summarizing and deciding with a language model, built by business users. For deterministic workflows that move records between systems at volume, the three tools above are the usual choice. The comparison with n8n is in Gumloop vs n8n.

Apps, code and the complexity ceiling

Zapier's directory is the largest by far, and for a niche SaaS tool it is the most likely to have a ready connector with authentication already solved. App count still misleads in two ways. First, an app being listed does not mean the trigger or action you need exists: check the specific operation. Second, n8n's HTTP Request node reaches any service with a documented API, so its smaller directory undersells what it can connect to, at the cost of an hour reading API docs that Zapier would have hidden.

On logic, the ceiling rises from Zapier to Make to n8n. Zapier handles linear workflows well and branches with Paths, but long multi-branch Zaps get hard to follow. Make handles branching and data reshaping visually without code. n8n treats code as normal: a Code node in JavaScript or Python at any point, sub-workflows you can reuse, and error workflows that run when another workflow fails. If you are new to n8n itself, start with what n8n is and what it takes to run.

Self-hosting and who holds the data

Zapier and Make run on the vendor's infrastructure. That is a real benefit: nothing to patch, back up or monitor. The trade-off is that every record your workflows touch passes through a third party, and your workflows only exist inside that vendor's account.

n8n is the only one of the three you can run on a server you control, which also lets workflows reach internal databases and systems that are not exposed to the internet. The cost is responsibility: someone updates n8n, watches the database and disk, rotates credentials, keeps backups and restores them when something breaks. If nobody on your side can own that, n8n Cloud gives you the same editor without the server, and Zapier or Make may be the more honest choice.

Before choosing a hosting model, answer four questions:

  • Does any contract, regulation or client policy restrict where customer data can be processed?
  • Do your workflows need to reach systems inside your network?
  • Who updates and monitors a self-hosted instance, and who covers them on leave?
  • Whose accounts hold the credentials: the company's, or one employee's?

AI steps and agents

All three can call a language model inside a workflow to classify an email, summarize a ticket or draft a reply. The differences show up with agents, where the model chooses which tools to call. n8n has agent nodes with memory, tool calling and vector store connections built into the canvas, and it can use self-hosted models when data cannot leave your servers. Zapier offers Zapier Agents, billed in "activities" separately from tasks (400 a month on the free plan, 1,500 on Pro, per its pricing page). Make offers AI modules and agents on its canvas; modules that run on Make's own AI provider can use more than one credit per operation.

For simple AI steps, any of the three is enough. For agents that touch money, customers or compliance, the deciding factor is control: where hard rules sit outside the model, what gets logged, and who approves an action before it runs.

What changes in production: limits, failures and the bill

Comparisons stop at features. What you live with is what happens when a workflow fails or a limit is reached.

  • Zapier at the task limit. With pay-per-task on, Zapier keeps running and bills extra tasks at 1.25 times the base rate on annual plans and 2.5 times on monthly plans, and stops at three times your subscription. With it off, Zaps stop until the next cycle. Allowances reset monthly, even on annual plans.
  • Make out of credits. Scenarios stop until credits are added. Incoming webhooks queue up to your plan's webhook queue size, and polling scenarios pick up from their last successful run. Unused credits expire at the end of the term.
  • n8n Cloud limits beyond executions. Starter allows 5 concurrent executions and runs of up to 5 minutes; Pro allows more concurrency and runs of up to 40 minutes. A burst of webhooks or a long batch job hits these before it hits the execution count.
  • Failures you do not see. On every platform, set an alert for failed runs and name an owner for it. A lead sync that fails silently costs weeks of leads before anyone notices, whatever the tool.
!
Price the platform on your three busiest workflows

Count the billed steps of each, multiply by expected monthly runs, and look the totals up on all three pricing pages. Then add who will maintain it and how many hours a month that takes. That second number is usually larger than the subscription.

Switching later: there is no import button

None of the three imports the others' workflows in a way worth trusting with live data. Moving from Zapier or Make to n8n means rebuilding each workflow, re-creating every credential and webhook URL, and running old and new side by side before cutting over. That is also the moment to merge three linear Zaps into one workflow and add the error handling they never had. The step-by-step is in how to migrate from Zapier or Make to n8n.

If you would rather have it done, 2V Automation runs Zapier and Make to n8n migrations: an audit of what you run today, a rebuild (not an import), a parallel run on real data, then the cutover, in accounts you own. A small set moves in a week or two; larger, complex accounts take longer.

Which one fits your team

If this describes youPickRead next
No developer, a few short workflows, need them running this weekZapiern8n vs Zapier
Ops team that builds its own, branching and line items, cloud is fineMaken8n vs Make
Thousands of runs a month, long workflows, the Zapier bill keeps risingn8n (Cloud or self-hosted)Migration guide
Data must stay on your servers, or workflows reach internal systemsSelf-hosted n8nSelf-hosted n8n
Mostly AI agents, built by business usersGumloop or n8nGumloop vs n8n
Some simple, some heavyZapier for the simple ones, n8n or Make for the heavy onesThis page, worked example

Running two tools is normal: a few Zaps a marketer owns next to the heavy workflows on n8n or Make costs less than forcing everything onto one platform. Before committing, rebuild one real workflow on your top two candidates' free plans and see who on your team can fix it when it breaks.

Every comparison of n8n, Make and Zapier

Frequently asked questions

Which is better, n8n or Make or Zapier?

None of them wins for every team. Zapier is the better pick when non-technical people will build and own short workflows. Make is better for branching, loops and data reshaping on a visual canvas. n8n is better when volume is high, data has to stay on your own servers, or the workflow needs code and AI agents, as long as someone technical owns it.

How much does n8n cost per month?

On n8n Cloud, Starter is $20 a month for 2,500 executions and Pro is $50 a month for 10,000 executions, both billed annually, per n8n's pricing page checked in November 2026. Business is $800 a month for 40,000 executions and is self-hosted. The Community edition is free to self-host: you pay for the server and the time to run it.

How many Zaps are free in Zapier?

Zapier's free plan is limited by tasks: 100 tasks a month, two-step Zaps only (one trigger, one action), one seat, and checks for new data every 15 minutes. Filters, Paths and premium apps need the Professional plan or higher.

What are n8n and Zapier?

Both are workflow automation platforms: you connect apps, pick a trigger, and the tool runs a series of steps each time it fires. Zapier is a cloud service with a guided, step-by-step editor and the largest app directory. n8n is a node-based workflow tool you can use as a cloud service or run on your own server, with code steps and AI agent nodes built in.

What is a good alternative to n8n?

If nobody on the team is technical, Zapier or Make: both are fully managed, and Make handles branching logic closest to n8n. If the work is mostly AI agents built by business users, Gumloop is the AI-first option. If you want n8n's control without running it yourself, n8n Cloud is the alternative to self-hosting.

Is Make cheaper than Zapier?

Usually, for multi-step workflows, but not because it bills fewer units. Make often counts more credits than Zapier counts tasks for the same workflow, since the trigger and searches use credits, while Zapier does not bill triggers, filters or formatting steps. What decides it is the price per unit on your plan, so count your real workflow on both pricing pages.

Find out what a week of your team’s time is worth.

One week inside your operations, $2,000 for a team of 1 to 20. You keep the map and the costed plan whether you build with us, with someone else, or not at all.