n8n vs Make.com 2026: price, self-hosting, GDPR and what to pick
Both tools do the same thing — connect your systems without programming. The difference is in how they bill, where they run, and what happens when you hit their limits.
Pick Make.com if you want to start quickly, there is nothing sensitive in your data, you do not want to look after a server and your scenarios are short. It is simpler and cheaper at the start.
Pick n8n if you need to keep data in your own infrastructure, you have long multi-step scenarios, or you will need custom code. The Community Edition can be hosted with no licence fee.
The one difference that decides most often: Make counts every module action as an operation, n8n counts one execution of the whole workflow regardless of the number of steps. On a twenty-step scenario that is a twentyfold difference in consumption.
How they bill — the main practical difference
This is the thing that surprises most companies only after the first month of running, so I put it first.
- Make.com, per its official documentation, counts every module action as a credit — adding a row to a spreadsheet, reading an email, transforming data. A five-module scenario spends five credits per run.
- n8n, per its official documentation, counts one execution of the whole workflow regardless of how many steps it contains and how much data it processes.
What that means in practice: if you process 1,000 orders a month and the scenario has 12 steps, in Make that is roughly 12,000 credits, in n8n 1,000 executions. On short scenarios the difference is negligible; on long ones it decides the price of the whole solution.
So before picking a tool it is worth sketching the scenario and counting the steps. That is five minutes of work that can save you repeatedly moving up a plan.
Prices per the official price lists
| Make.com | n8n | |
|---|---|---|
| Free plan | yes — up to 1,000 credits a month | Community Edition self-hosted (no licence fee) |
| Lowest paid plan | Core from $9 / month | Starter from €20 / month on annual billing, 2,500 executions |
| Next plan | Pro from $16 / month | Pro from €50 / month on annual billing, 10,000 executions |
| Team plan | Teams from $29 / month | Business from €667 / month on annual billing, 40,000 executions |
| Enterprise | individual | individual |
| Self-hosting | not available | yes (Community Edition and higher plans) |
| Unit of consumption | credit = module action | execution = run of a whole workflow |
Prices per the official make.com and n8n.io price lists, checked 5 September 2026. Make quotes in dollars, n8n in euros. Price lists change — verify them directly at the sources below before deciding.
Comparing the numbers directly is misleading, because the units are not the same. A meaningful comparison only comes once you count the steps in your scenario and multiply by the expected number of runs.
Self-hosting: a difference that cannot be closed
Make is a cloud service and has no self-hosted version — it never has. n8n has a Community Edition that can be run on your own server with no licence fee.
For companies with sensitive data that is often the single deciding criterion. Health, payroll or client data may simply not be permissible to send to a third-party cloud service.
But self-hosting is not free. Free of licence fees, yes; the server, backups, updates and administration time cost money. At low volume the cloud is usually cheaper on total cost. A detailed budget is in Self-hosted n8n for business.
Comparison by criteria
| Criterion | Make.com | n8n |
|---|---|---|
| Ease for a beginner | higher — the visual editor is friendlier | steeper learning curve |
| Speed of a first scenario | very fast | fast, if an instance is already running |
| Number of ready integrations | a very broad app catalogue | a broad catalogue + a universal HTTP node |
| Custom code | limited options | a JavaScript node, practically unrestricted |
| Working with AI models | supported | supported, including nodes for vector databases |
| Where the data sits | with the provider (AWS) | wherever you decide |
| Portability of the solution | tied to the platform | workflows export as JSON |
| Server administration | none | with self-hosting, on you or your supplier |
| Suits | a fast start, short scenarios, no IT | sensitive data, long scenarios, a need for control |
GDPR: what self-hosting solves and what it does not
Comparisons simplify this in a way that can be dangerous, so more precisely:
- Self-hosted n8n removes one processor — data handled in a workflow does not pass through the automation platform provider's infrastructure.
- It does not cover external calls. If a workflow sends data to a language model or another service, it goes there regardless of where n8n runs.
- Make hosts in both the EU and North America per the information on its site; the specific location and contractual terms should be verified when signing.
- GDPR compliance is not determined by the tool but by the lawfulness of the processing, information obligations and processor agreements.
Nexflow does not provide legal advice. This is a technical description of where data travels; assessing compliance belongs to your DPO or lawyer.
How to decide in five minutes
- Is there sensitive personal data involved? If yes and it must not go to a third-party cloud → self-hosted n8n. You can stop reading.
- How many steps does your typical scenario have? Above roughly ten steps at higher volume, per-module-action billing starts to bite → n8n.
- Will you need custom code or an integration that does not exist? If yes → n8n.
- Is there anyone to administer a server? If not, and the previous points do not apply → Make.com.
- If you did not stop anywhere → Make.com. It is simpler and you can change later.
The recommendation I give most often: start in Make, prove on one process that automation makes sense, and move to n8n only when there is a substantive reason — volume, sensitive data or a need for custom code. Migrating is not free, but it is cheaper than months of delay while “the right tool” is being chosen.
What does not decide the choice
The number of integrations in the catalogue. Both have more than you will use. What matters is whether they have yours — and if not, whether an HTTP call solves it.
Which one is “more modern”. Both are developing and both handle language models.
What somebody else uses. Your situation decides — volume, data, and who will administer it.
Trying to pick a tool “forever”. Workflows can be rewritten. Delaying your first automation by six months costs more than a later migration.
When you do not need Make or n8n
- The connection already exists. If your e-shop, accounting system or CRM has the integration you need built in or as a ready-made add-on, that is simpler than another tool you have to pay for and maintain.
- The process runs a few times a month. If it takes a few minutes, doing it by hand is cheaper than setting up and maintaining a scenario.
- The process is still changing. Let it settle first. A workflow built on something that changes next month will need rework in either tool.
- No one will watch the automation. When an API changes or access expires, the scenario stops. If no one gets the error alert and acts on it, that is a risk, not a saving.
Frequently asked questions
Which is cheaper, n8n or Make.com?
It depends on scenario length and volume. Make counts every module action, n8n one execution of a whole workflow — so on long scenarios n8n is markedly cheaper. On short scenarios at low volume Make tends to be cheaper, especially once you count server administration into n8n's cost.
Does Make.com have a self-hosted version?
No. Make is a cloud service that has never offered a self-hosted version. Its site mentions an on-prem agent for reaching a local network in enterprise deployments, but that is not the same as running the whole platform yourself.
How are operations counted in Make and in n8n?
Per its official documentation, Make counts every module action as a credit — a five-step scenario spends five credits per run. n8n counts one execution of the whole workflow regardless of the number of steps and the volume of data.
Which is easier to start with?
Make.com. The visual editor is friendlier and you will build a first scenario faster. n8n has a steeper learning curve but offers more control — custom code, self-hosting and portability of the solution.
Is n8n better for GDPR?
When self-hosted it removes one processor, because data does not pass through the platform provider's infrastructure. It does not cover external calls, though — if a workflow sends data to a language model, it goes there either way. GDPR compliance itself is determined by the lawfulness of the processing and by contracts, not by the choice of tool.
Can we move from Make to n8n?
Yes, but not automatically — the scenarios have to be rebuilt, since the platforms share no common format. That is why I recommend deciding up front for long scenarios and high volume, where the billing difference is fundamental.
Which should a small business choose?
If there is nothing sensitive in your data, the scenarios are short and there is nobody to administer a server — Make.com. If any of the opposite applies, n8n. And if you are unsure, start in Make: you will find out whether automation pays off for you at all, and you can move later.
Sources
- Make.com — official price list — checked 5 September 2026
- n8n — official price list — checked 5 September 2026
- n8n — official documentation — definition of a workflow execution, self-hosting
- Make.com — official documentation — definition of an operation/credit
Related services
Related articles
Rather not decide blind?
Describe the process you want to automate and the rough volume. I will tell you which tool suits you better and what running it will cost — even if that means you build it yourself.
Get a free consultation →