n8n vs Make vs Zapier: Which Should You Choose?
All three tools do the same job under three different economies: n8n bills per workflow run, Make and Zapier bill per step. Choose n8n if your workflows run past five or six steps or your data must stay on your own infrastructure; Make for short, low-volume workflows; Zapier for the widest app coverage and the fastest start.
At a glance
n8n
You pay per run, and the number of steps in a workflow does not change the price. It can run on your own servers, and every plan includes unlimited users. The cost: someone has to operate that server, and it has the smallest catalogue of ready integrations of the three.
Make
You spend a credit for every module action. Its visual editor is the clearest of the three and it is the cheapest option for short workflows. The cost: there is no self-hosting option, and scenarios stop when credits run out.
Zapier
You spend a task for every action step. With over 9,000 apps its coverage is far ahead, and you can ship your first automation the same day. The cost: data is processed in the United States, and its contract prohibits processing sensitive data.
Side-by-side comparison
| Criterion | n8n | Make | Zapier |
|---|---|---|---|
| Billing unit | Execution: one full run of the workflow. Step count is irrelevant. | Credit: one per module action. Router and error-handler modules are not counted. | Task: each successful action step. Triggers, Filter, and Formatter are not counted. |
| Does step count raise the bill | No. A 3-step and a 40-step workflow are both one execution. | Yes, linearly. | Yes, linearly. |
| Published entry price | Starter $20/mo: 2,500 executions. Pro $50/mo: 10,000 executions. (Billed annually) | Core $12/mo: 10,000 credits. Pro $21/mo, Teams $38/mo at the same credit tier. | Professional from $19.99/mo, Team from $69/mo. Price rises with the task tier you pick. |
| Free tier | Community Edition: free and unmetered on your own server. | 1,000 credits/month, 2 active scenarios, 15-minute minimum interval. | 100 tasks/month, two-step Zaps only, 15-minute polling. |
| Run it on your own servers | Yes. SSO, Git version control, and environment separation come with Business ($800/mo). | No. | No. |
| Where data is processed | Cloud: EU, Frankfurt. Self-hosted: wherever you deploy it. | AWS, EU or North America region. | United States. The DPA names Zapier Inc., San Francisco, as data importer. |
| Sensitive data (health, biometric, criminal records) | No contractual restriction. Self-hosted, the data never leaves your infrastructure. | No prohibition in the DPA; standard controller-processor terms apply. | Explicitly prohibited: the service is not designed for it and you agree not to submit it. |
| Ready integrations | 2,012 integrations. Anything missing is covered by writing a custom node. | 3,000+ apps. | 9,000+ apps. The widest coverage of the three. |
| Users | Unlimited on every plan, including the $20/mo Starter. | Teams and team roles from the Teams plan onward. | One seat on Free and Professional. 25 users from Team (from $69/mo). |
| Writing code | JavaScript and Python steps on every plan; bash and custom nodes when self-hosted. | Make Code App runs JavaScript or Python; 2 credits per second of execution time. | Code by Zapier; 30 seconds of runtime included per step on Professional and Team. |
| Licence | Sustainable Use License (fair-code). Source-available, commercial resale restricted. | Closed-source SaaS. | Closed-source SaaS. |
| What happens at quota | Workflows keep running; on Business, overage is invoiced. | Scenarios stop until you buy extra credits or move up a tier. | Continues at a higher per-task rate if pay-per-task is on; otherwise Zaps pause. |
| Fastest trigger interval | No platform limit; you set the schedule. | 1 minute on paid plans, 15 minutes on free. | 2 minutes on Professional, 1 minute on Team and Enterprise. |
The same workflow, priced three ways
The billing-unit difference stays abstract until you price the same workflow three ways. Each scenario below is a single workflow running 1,000 times a month; the only variable is how many steps it contains.
| Workflow | n8n | Make | Zapier |
|---|---|---|---|
| 3-step workflow, 1,000 runs/month | 1,000 executions | ~3,000 credits | ~3,000 tasks |
| 12-step order workflow, 1,000 runs/month | 1,000 executions | ~12,000 credits | ~12,000 tasks |
| 40-step cross-department workflow, 1,000 runs/month | 1,000 executions | ~40,000 credits | ~40,000 tasks |
- The n8n column never moves: execution count is independent of how many steps the workflow contains. The Make and Zapier columns grow linearly with step count.
- Make does not count router or error-handler modules, and Zapier does not count the trigger or Filter/Formatter steps, so real consumption lands slightly below these figures.
- The reverse holds too, and it favours Make: on a 3-step workflow, Make Core gives you 10,000 credits for $12/month — roughly 3,300 runs — against 2,500 runs for $20/month on n8n Starter. For short workflows Make is both cheaper and roomier.
- The crossover sits at roughly five to six steps. Below it Make wins; above it n8n pulls ahead, and the gap widens with every step you add.
- 12,000 credits and 12,000 tasks both exceed the tier the entry prices above refer to. Make and Zapier price in volume tiers, so use the tier selector on their own pricing pages for what this scenario costs today.
Which one, and when
Choose n8n when
- Data has to stay on your own infrastructure: healthcare, finance, public sector, or anywhere GDPR obligations dominate the decision.
- Your workflows are long. Past five or six steps per run, per-step billing starts working against you.
- Volume is high or unpredictable; a fixed execution quota is easier to manage than a step bill that moves with your traffic.
- Someone on the team writes JavaScript or Python, or you need a custom node for a system nobody has integrated yet.
- Several people will touch the automations: n8n includes unlimited users on every plan.
- You are building AI agents or a RAG system; n8n's agent and vector-store nodes are built for exactly this.
Choose Make when
- An operations team with no developer will build the workflows themselves; Make's visual editor is the most readable of the three.
- Workflows are short (two to four steps) and volume is low. In that band Make is the cheapest of the three.
- You do not want to operate a server, and keeping data in the EU region is sufficient for your obligations.
- The workflow diagram itself is part of how the team communicates; Make's visualisation serves that best.
Choose Zapier when
- The app you need to connect is niche and the other two do not have it. With 9,000+ integrations, that risk is lowest here.
- You need to start and finish today, and a two-step workflow solves your problem.
- One person will build and own the automation, so Professional's single seat is not a constraint.
- You want to start small inside your own department without waiting on an enterprise IT process.
Reconsider all three when
- You will process health data, biometric data, or criminal records: Zapier's own contract rules this out, and the work belongs on a self-hosted deployment or in purpose-built software.
- The automation will become part of your product and be sold to your customers: n8n's licence does not permit white-label resale, and Make's and Zapier's per-step pricing compounds badly in that model.
Frequently Asked Questions
Which one is cheapest?
It depends on how many steps your workflow has, and the crossover sits at roughly five to six. On a 3-step workflow, Make Core gives about 3,300 runs for $12/month against 2,500 runs for $20/month on n8n Starter, so Make is both cheaper and roomier. On a 40-step workflow, the same 1,000 runs are still 1,000 executions on n8n but 40,000 credits on Make. Make wins short, n8n wins long, and the gap widens with every step you add.
Is n8n open source?
Not quite, and the distinction matters. n8n uses a fair-code licence called the Sustainable Use License: the source is public, and you may run and modify it on your own server for free. The restriction sits in one place — you may not repackage n8n as your own product and sell it, or host it and charge people for access. n8n's own documentation cites the Open Source Initiative definition and states it does not call itself open source. Consulting on n8n, building workflows for clients, and writing custom nodes are all explicitly permitted. Make and Zapier offer no equivalent: both are closed-source SaaS.
Can we run n8n on our own servers for free?
Yes. Community Edition runs free on your own infrastructure with no cap on workflows. Three things enterprises usually miss are absent from it: SSO and LDAP, Git version control, and separated development and production environments. Those unlock on the Business plan ($800/month). The hidden cost of free is the server itself — hosting, backups, upgrades, and monitoring land on someone. Infrastructure for a self-hosted n8n instance runs roughly $5 to $40 a month; the real cost is not the money but the time keeping it healthy.
Where is our data processed, and does it matter for GDPR?
It matters a great deal. Zapier processes your data in the United States; its DPA names San Francisco-based Zapier Inc. as the data importer and relies on the EU-US Data Privacy Framework, with Standard Contractual Clauses as the fallback. Make hosts on AWS in either an EU or a North America region, and its DPA follows the same framework-then-SCC structure. n8n's cloud runs in Frankfurt, and a self-hosted deployment sits wherever you put it. If your compliance position is that personal data should not leave your own infrastructure at all, self-hosted n8n is the only one of the three that satisfies it, because there is no transfer to assess.
Which should we choose if we process health or other sensitive data?
Zapier is ruled out: its contract states the service is not designed for sensitive information and you agree not to submit any. That definition covers health data, biometric data, criminal records, and categories such as racial origin, political opinion, and religious belief. Make's DPA carries no such prohibition; standard controller-processor terms apply. n8n imposes no contractual restriction, and on a self-hosted deployment the question does not arise because the data never leaves your infrastructure. For teams in regulated sectors we recommend self-hosted n8n in practice.
How long does migrating from Zapier or Make to n8n take?
There is no automatic converter; workflows get rebuilt. Our own measurements land around one to two weeks for ten to twenty simple Zaps or scenarios, and two to four weeks for a mid-complexity system carrying real business logic. The driver is not the number of workflows but how well the business rules inside them are documented. We normally migrate incrementally, starting with the highest-volume workflow, so the cost reduction shows up in week one and the old system stays available as a fallback for a while.
How many people can access the automations?
The gap here is larger than people expect. n8n includes unlimited users on every plan, including the $20/month Starter. Zapier gives one seat on Free and Professional; team access means moving to the Team plan (from $69/month) and its 25 users. On Make, teams and team roles unlock from the Teams plan onward. If more than one department will touch the automations, this line can matter more than the execution price.
You are an n8n Ambassador — isn't this comparison biased?
To keep it from being biased, we did not generate the numbers: every figure on this page comes from the vendors' own published pages, and the sources are listed below. The Ambassador credential gives us an advantage in knowing n8n well, not an obligation to recommend it everywhere. This page in fact declines to recommend n8n in three places: Make is cheaper for short, low-volume workflows, Zapier's app coverage is unmatched for niche integrations, and n8n's licence does not permit selling the automation as your own product. A comparison is only useful once it shows where its author's tool loses.
Sources
- n8n plan prices, execution definition, quotas, and concurrency limits — n8n.io/pricing
- n8n integration count (2,012) — n8n.io/integrations
- n8n Sustainable Use License and the fair-code model — docs.n8n.io
- Make plan prices, credit definition, app count, and hosting region — make.com/en/pricing
- Make data processing terms and international transfer conditions — Celonis DPA for Make
- Zapier plan prices, task definition, seat counts, and app count — zapier.com/pricing
- Zapier sensitive-data prohibition and US transfer terms — Zapier Data Processing Addendum
Prices and limits were read from the vendors' own pages on 7 August 2026.
Related service
Our n8n workflow automation service and price bandsDisclosure: humaix founder Onur Polat is Türkiye's official n8n Ambassador. Every figure on this page was taken from the three vendors' own published pages and sourced below, and the cases where Make or Zapier is the better choice are stated with the same directness.
Last updated: 7 August 2026
Talk Through Which One Fits You