n8n vs Zapier vs Make: which automation tool fits your company?

Every price and rule on this page comes from the vendors’ own pricing, help and legal pages, read on October 1, 2026. Prices change often, so check the vendor’s page before you sign up. We build workflows on all three tools, which is why this guide gives you criteria and a worked example, and names no winner.

n8n vs Zapier vs Make, side by side

The main differences are the unit you pay for, where your data can live and whether you can run the tool yourself. Everything else follows from those three.

Two lines of this table decide most choices. If your data must be stored only in the EU, Zapier’s own privacy page says it does not offer that option, while Make and n8n Cloud do. If you want the tool on your own infrastructure, n8n is the one of the three that offers it.

How each tool counts what you pay for

Zapier counts actions, Make counts module operations and n8n counts whole runs. The same workflow therefore costs very different amounts from one tool to the next, and the gap grows with the number of steps.

Zapier: tasks. Zapier’s help center defines a task as “any successful action that runs in Zapier”. Trigger steps do not count, nor do Filter and Paths steps, Formatter, Delay, Looping and a few other built-in apps. Failed actions do not count either. Replaying a whole Zap run counts its successful steps again. Some products have their own rates: AI by Zapier depends on the model tier you pick, and each successful tool call through Zapier MCP counts as 2 tasks.

Make: credits. Make’s pricing page says that each module action in a scenario, “like adding a Google Sheet row or fetching Gmail account data, counts as one credit”, and that some features using Make’s AI Provider “may use more”. Its FAQ lists what consumes credits: reading data from an app or a webhook, searching, creating, updating, deleting, transforming, aggregating and iterating. Routers and error handler modules consume none. When you run out, “your scenarios won’t continue to run” until credits are added, with warnings at 75 and 90 percent.

n8n: executions. n8n’s pricing page states that “an execution is a single run of your entire workflow”, whatever the number of steps or the amount of data it processes. On the Business plan, going over the quota does not stop workflows; overage is billed at 4,000 euros per extra bucket of 300,000 executions if you do not move up a tier.

A worked example: one workflow, three bills

Take one workflow with a trigger and five actions, running 2,000 times a month: a web form arrives, then the workflow creates the contact in the CRM, creates a deal, adds a spreadsheet row, posts a message to the sales channel and sends a confirmation email. None of the steps is a search, which Zapier counts by its own rule.

The arithmetic is ours; the prices are the vendors’. Read the result with care. Filters, branches, loops and AI steps change the count on every tool, and each plan has limits beyond billing: n8n’s Starter plan allows 5 concurrent executions and a maximum execution time of 5 minutes, and Zapier checks for new data every 2 minutes on Professional and every 15 minutes on Free. Self-hosting n8n removes the n8n bill but adds a server and someone to look after it, which this table does not show.

Our ROI calculator helps put the other side of the equation into figures: the hours the workflow gives back.

Where your data goes

For a European company, data location often settles the question before price does. Each vendor states it on its own pages.

n8n’s pricing page says that for hosted plans “data is stored within the EU, on servers located in Frankfurt, Germany”, and for self-hosted plans “wherever you decide to host n8n”.

Make’s help center explains that each organization can select a data center in the United States or in the European Union, and that “you cannot change the location of the data center after you create the organization”. Pick the EU region when you create the organization.

To the question “Is there an option to have my data stored only within the EU?”, Zapier’s data privacy page answers: “Zapier does not support this option.” For transfers from the EU, it relies on its certification under the EU-US Data Privacy Framework, which the page ties to the European Commission’s adequacy decision of July 2023.

None of this replaces your own review. Check the data processing agreement, the list of subprocessors and what the workflow actually carries: an automation that only moves order numbers is not the same risk as one that copies customer files.

The n8n license, in plain terms

n8n is not open source in the strict sense. It is published under the Sustainable Use License, which n8n describes as a “fair-code” license it created in 2022.

The license lets you use, modify and redistribute n8n with three limits, in n8n’s words: use “only for your own internal business purposes or for non-commercial or personal use”, distribution only “free of charge for non-commercial purposes”, and no removal of licensing or copyright notices. Files marked .ee. fall under a separate n8n Enterprise License.

n8n’s license FAQ gives concrete cases. Allowed under the free license:

• building automations for your clients on your instance, “as long as your clients do not have the possibility to create or edit them”;
• charging clients for workflow creation, setup, maintenance, training and consulting;
• installing, configuring and maintaining n8n on a client’s own server, provided you are not also hosting it for them.

Not allowed without a commercial agreement: hosting n8n as a service where clients build their own workflows, letting external users configure workflows through your product, forking n8n to launch your own automation product, or white-labeling it. When a case is unclear, n8n asks you to write to [email protected].

For a company using n8n for itself, none of this gets in the way. It matters when a provider offers to host n8n for you: ask how the setup fits these rules. When a project calls for a paid license, such as n8n’s Business or Enterprise plan or a commercial agreement, we take it and put it in the quote. Our guide to n8n self-hosted goes through each license case and what a production setup needs.

Which tool fits which company

There is no best tool in the abstract. Match the tool to your volume, your data rules and the people who will maintain the workflows.

Zapier fits a team without technical staff that needs to connect many cloud apps with short, linear flows, and whose data can be processed in the United States. It is the quickest to start with.

Make fits a team that wants visual scenarios with branches and data transformation, at a lower entry price, and needs the data in the EU. Keep an eye on credit consumption: polling triggers and loops use credits fast.

n8n fits long workflows where per-step billing would climb, data that must stay on your infrastructure, and teams with someone at ease with APIs. Its AI agent nodes and code steps make it a natural home for workflows that include AI, as described in our guide on how to build an AI agent. Our guide to the n8n AI agent shows how its agent node, tools and human approval work.

None of the three fits when the process needs a real application with its own screens and rules, or when volumes make any per-run pricing absurd. Then the workflow belongs in code, next to your systems. And when a step has to happen on the screen of an application with no API, our guide RPA vs AI agents helps you decide between a robot and an agent.

Mixed setups are common. A company can keep Zapier for a handful of marketing flows, and run its supplier invoice processing or its order intake on n8n. Our n8n agency sets up and runs n8n in your name, and our Make and Zapier team audits and repairs existing flows before anyone talks about moving them. To see what an AI step inside such a workflow does with a messy document, try the document extraction demo.

What to check before you commit

The tool matters less than what happens the day a workflow breaks. Before you choose, and before you sign with anyone who builds for you, check four things.

1. Error handling. n8n offers error workflows and automatic retries. On Zapier, failed actions do not count as tasks, but they still have to reach someone. On Make, error handler modules consume no credits, and running out of credits stops scenarios. Whatever the tool, each workflow needs an alert that reaches a named person.
2. Ownership. The account, the API keys and the workflows should be in your company’s name, with access for the builder that you can revoke.
3. Documentation. What triggers each flow, what it changes and what to do when it fails, written for the person who will take over.
4. The bill at twice the volume. Recount tasks, credits or executions with your own figures, doubled, and see which plan you land on.

If your team will maintain the workflows itself, n8n training on your own instance is often the fastest way to get there. For the wider picture, from the first process to automate to AI steps, see our AI automation and process automation pages.

Get your workflows running without the guesswork

The hours come back once the workflows run every day, and someone hears about it the moment one fails. In a free 30-minute assessment, bring the process you want to automate and the tools you already pay for. We tell you which of the three fits, what it would cost at your volume, and what to build first. Every workflow we deliver ships with an error alert, documentation and accounts in your name, and the scope and price are fixed in writing before we start.

Book your free 30-minute assessment, or put a figure on the hours you would get back with the ROI calculator. We reply within one business day.