Best European Alternatives to Zapier
Looking for a European alternative to Zapier? An automation platform holds credentials to every system it connects — CRM, email, payments, database — and Zapier is a US processor holding all of them, on per-task pricing that scales badly.
Europe leads this category outright: the best visual automation builder and the best self-hostable one are both European, and most options here can run on your own infrastructure.
How we rank these tools — 4-step process
-
1
European ownership, verified
The company is headquartered and incorporated in the EU, EEA or Switzerland, and processes customer data in Europe. A US parent company disqualifies a tool from this page regardless of where its servers are.
-
2
Category fit and hands-on review
What the tool actually does, who it suits, and where it falls short — checked against the vendor’s own documentation, changelog and pricing page rather than its marketing copy.
-
3
Compliance and pricing check
GDPR posture, hosting location and the prices quoted on this page are verified against the vendor’s public pricing before publication, and re-checked when we revisit the category.
-
4
Position on this page
Placement on this page can be paid, and that can affect which tools appear here and the order they appear in. It never buys a good review: a tool that fails the checks above is not here at any price, and payment does not change the shortcomings we write about. A vendor can ask us to correct a factual error — not to remove a criticism.
Vendors can pay for visibility on this page. It never changes what an entry says about a product, including the criticism, and we earn nothing when you click through to a vendor. Paid placement can affect which tools appear here and the order they appear in. Editorial policy
7 European Alternatives to Zapier
n8n
Source-available workflow automation you can fully self-host
Make
Visual scenario builder with more than 2,000 integrations
Kestra
Declarative orchestration for data and infrastructure workflows
Node-RED
Flow-based wiring for APIs, services and hardware
Locoia
Integration and automation platform for mid-market operations
Automatisch
Open-source Zapier alternative, self-hosted by design
elastic.io
Embeddable integration platform for software vendors
Key takeaways
- Zapier holds live authorised connections to more than 9,000 applications, so the exposure is your credentials rather than your data.
- Self-hosting n8n, Automatisch, Kestra or Node-RED is the only move that puts those tokens on a machine you own; switching to a European cloud relocates them.
- The free plan is 100 tasks a month and two-step workflows only; Professional runs from $19.99 to $3,389 a month billed annually.
- Zapier counts every successful action, so a well-structured workflow with filters and branches costs more per run than a careless one.
- Make is European and does not change the arrangement: it is a hosted platform holding the same credentials, with the data centre chosen at signup.
Why people leave Zapier
Zapier is the best-connected automation tool in the world and nothing on this page beats it on catalogue. More than 9,000 applications, a builder anyone can learn in a morning, and an answer already written for whatever you are trying to connect.
The reason to look elsewhere is not a feature and not a price. It is that the thing Zapier holds is not your data — it is your credentials.
To move a lead from a form into a CRM, Zapier keeps an authorised connection to both, and after two years of accumulation that means one company in San Francisco holds live access to your CRM, your accounting system, your mailbox, your invoicing and your file storage. No individual supplier in that list would be given that access on its own.
The second reason is what happens to the bill as the automations grow. The free plan allows 100 tasks a month and only two-step workflows; Professional runs from $19.99 a month billed annually to $3,389 at the top of the range, for 750 tasks up to two million. Every successful action counts, so a branched workflow consumes several per run, and the invoice grows with the thing you were trying to make routine.
- One supplier holds the keys to everything else An automation platform works by keeping authorised connections open, and the authorisations are the product. Revoking Zapier does not merely stop the workflows; it is also the only way to withdraw a standing permission that has been quietly widening since whoever set it up left the company. Self-hosting changes the arrangement rather than relocating it: the tokens sit on a server you own, and nobody else can be compelled to produce them because nobody else has them.
- Task pricing charges you for reliability Zapier counts every successful action, so the well-built workflow with a filter, a lookup and two branches costs several tasks each time it runs, while the sloppy one that does everything in a single step costs one. That is a pricing model that quietly rewards the wrong engineering. The tools here that charge per deployment or per server do not care how many steps a workflow has.
- The cheap plans are deliberately thin Multi-step workflows require a paid plan; the free tier is limited to two steps and 100 tasks a month. Code steps are capped by runtime, from one second on Free to two minutes on Enterprise, so a transformation that takes a moment too long has nowhere to go inside the product. Autoreplay for failed steps starts at Professional. None of these is unreasonable as a business model; all of them are absent when the engine runs on your own hardware.
- No published European region Zapier, Inc. operates from San Francisco, its data processing addendum points at standard contractual clauses, and there is no European region documented on its pricing or plan pages. That is a lawful arrangement and a perfectly ordinary one. It is also a different answer from the one several tools here give, where the question does not arise because the engine is on your infrastructure.
What you have to replace, not just match
Before comparing anything, print your Zap list and sort it into three piles, because they are three different products and this page contains all three.
The first pile is app-to-app glue reacting to events: a form submission creates a CRM record, a won deal posts to a channel. That is what Zapier is, and n8n, Make, Automatisch and Locoia are the replacements.
The second pile is scheduled data work with dependencies, retries and the need to rerun yesterday because the source was wrong — that is orchestration, and it is Kestra's territory, not Zapier's. The third pile is anything touching hardware or an industrial protocol, which is Node-RED.
elastic.io belongs to none of the three. It is bought by software companies that need to ship integrations to their own customers inside their own product, which is a different purchase from automating your own back office and is on this page because a surprising number of Zapier accounts are a software vendor doing it the hard way.
The alternatives compared
| Position | Tool | Headquarters | Pricing | Jurisdiction |
|---|---|---|---|---|
| #1 | n8n | Berlin, Germany | Free self-hosted / Cloud from about €24/month | EU (Germany) |
| #2 | Make | Prague, Czech Republic | Free (1,000 operations/month) / Core from about $9/month | EU (Czech Republic) |
| #3 | Kestra | Lille, France | Free and open source / Enterprise on request | EU (France) |
| #4 | Node-RED | Hursley, United Kingdom | Free and open source | Wherever you run it |
| #5 | Locoia | Berlin, Germany | From about €200/month; embedded and enterprise on request | EU (Germany) |
| #6 | Automatisch | Berlin, Germany | Free self-hosted / hosted cloud from about €20/month | EU (Germany) |
| #7 | elastic.io | Bonn, Germany | Custom pricing by deployment and volume | EU (Germany) |
How each alternative compares to Zapier
- Which law reaches it. EU (Germany). Zapier is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Your own servers, or EU cloud.
- Source code. Open source, where Zapier is not: you can read what it does rather than take the description on trust.
- Independently checked. Source-available (Sustainable Use Licence).
Best for: Teams automating across systems whose access they cannot hand to a third party
n8n GmbH is in Berlin, and the Community Edition runs on your own servers for free with no task meter of any kind. That is the direct answer to both of Zapier's problems at once: the tokens for your CRM, your accounting package and your mailbox sit on a machine you own, and a workflow firing every minute costs exactly what one firing daily costs.
It is also considerably more capable where Zapier is thinnest. Real JavaScript and Python code nodes take over when the visual builder runs out, without the runtime ceiling Zapier applies — one second on Free, two minutes at Enterprise — because it is your server deciding how long something may run.
Branching, looping, error workflows, webhooks and queue mode cover the operational shape of real automation, and native AI agent and LangChain nodes make an agent a first-class step rather than an HTTP call.
Two things to know before committing. Self-hosting is genuine operational work: a server, a queue, upgrades and somebody who notices a failure at two in the morning.
And the Sustainable Use Licence is source-available rather than open source, permitting free internal use while restricting offering n8n as a service to others — fine for automating your own business, worth reading closely if you intend to build a product on it. The EU cloud starts from about €24 a month for teams who want the jurisdiction without the server.
What n8n does better than Zapier
- Self-hosted, so the authorisations to every connected system never leave infrastructure you control
- No task meter at all when self-hosted, where Zapier charges for every successful action
- JavaScript and Python code nodes with no runtime ceiling, against Zapier's one second to two minutes
- German company with an EU cloud option for teams that do not want to run a server
- Native AI agent and LangChain nodes rather than an HTTP call to somebody else's API
Where n8n is a step down from Zapier
- Around 400 integrations against Zapier's 9,000-plus, so the long tail is where it hurts
- A steeper learning curve, and a builder aimed at technical users rather than at anyone
- Self-hosting is real operational work with a real person attached to it
- The Sustainable Use Licence is source-available, not open source, which matters if you build on it
Standout against Zapier. It is the only option here that combines a self-hosted engine with unrestricted code steps, so the difficult transformation runs on your hardware under your rules.
- Which law reaches it. EU (Czech Republic). Zapier is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. EU or US data centre, chosen at signup.
- Source code. Closed source, as Zapier is.
Best for: Teams who want branching and an EU data centre without running anything
Make is in Prague, part of Celonis, with more than 2,000 integrations plus custom API modules, and you pick an EU or US data centre when you sign up. Choosing EU puts both the processing and the stored credentials inside European jurisdiction, which is a real improvement on where they are now.
The reason to choose it over Zapier is expressive power rather than geography. Routers branch, iterators walk an array item by item, aggregators collapse results back together, and error handlers decide what happens when a step fails. A real business process has all four of those shapes, and a linear builder forces you to fake them with duplicated workflows that drift apart. The visual debugging is genuinely better too, showing exactly what entered and left each module.
What it does not do is change the arrangement this page is about. There is no self-hosting, so Make holds the same standing authorisations Zapier does; the difference is the country and the company, not the custody. Pricing is per operation, so a branched scenario consumes several per run, with a free tier of 1,000 operations a month and Core from about $9.
What Make does better than Zapier
- Routers, iterators and aggregators express branching logic Zapier cannot, without duplicating workflows
- An EU data centre chosen at signup, so processing and stored credentials sit in Europe
- More than 2,000 integrations plus custom API modules for anything unsupported
- Much cheaper per operation than Zapier, with 1,000 free operations a month
- Visual debugging that shows exactly what each module received and returned
Where Make is a step down from Zapier
- No self-hosting, so the vendor still holds authorised access to every connected system
- Operation-based pricing, so complexity costs money on every run exactly as it does on Zapier
- Owned by Celonis rather than independent, which is a different risk from being American
- Complex scenarios have a real learning curve that Zapier's linear builder avoids
Standout against Zapier. It is the only hosted platform here that can branch three ways and loop through an array without duplicating the workflow — and it is honest about still holding your keys.
- Which law reaches it. EU (France). Zapier is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Your own infrastructure, or EU cloud.
- Source code. Open source, where Zapier is not: you can read what it does rather than take the description on trust.
Best for: Teams whose scheduled workflows keep needing yesterday rerun
Kestra Technologies is in Lille, Apache 2.0 licensed and free, and it is a data orchestrator rather than an app-integration tool. Confusing the two is the most expensive mistake in this category, so it is worth stating up front: Kestra does not replace Zapier, and if all your Zaps are "new form submission creates a CRM record", this is the wrong entry on the page.
It replaces the other pile. A surprising number of Zapier accounts contain scheduled workflows pulling a file, transforming it and loading it somewhere, held together with a daily trigger and hope — and those are pipelines pretending to be Zaps. Kestra gives them what they actually need: dependencies, retries, backfills and replays, so rebuilding yesterday's run because the source system was wrong is a supported operation rather than a manual afternoon.
Workflows are declarative YAML instead of Python, which means a pipeline change arrives as a reviewable diff rather than as code somebody has to read, while tasks themselves can run scripts in any language. Six hundred plus plugins cover the usual data systems. The limits are honest: not suited to app-to-app integration, a younger and smaller community than Airflow's, and self-hosting needs a database and a queue.
What Kestra does better than Zapier
- Backfills and replays, so rerunning a failed day is a supported operation rather than manual repair
- Dependencies and retries as first-class concepts, which a linear Zap has no way to express
- Declarative YAML workflows that can be reviewed in a pull request and generated by tooling
- Apache 2.0 and free on your own infrastructure, with a French company behind it
- Tasks can run scripts in any language, with no runtime ceiling imposed by a plan tier
Where Kestra is a step down from Zapier
- Not an app-integration tool, so most Zaps have no equivalent here at all
- Assumes a technical team: YAML, a database and a queue are the entry requirements
- A younger, smaller community than Airflow, let alone Zapier
- Enterprise features are quote-priced rather than published
Standout against Zapier. It is the only tool here that can rebuild last Tuesday because the source system was wrong, which is the failure mode a scheduled Zap handles by silently producing the wrong answer.
- Which law reaches it. Wherever you run it. Zapier is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Wherever you run it.
- Source code. Open source, where Zapier is not: you can read what it does rather than take the description on trust.
Best for: Factory floors, building automation and anything with a sensor in it
Node-RED is an OpenJS Foundation project that originated at IBM in the United Kingdom, Apache 2.0 licensed and completely free, and it automates physical things. MQTT and Modbus are native, the whole thing fits on a Raspberry Pi or an industrial gateway, and sensors, controllers, APIs and dashboards get wired together in a visual editor.
That is not a comparison with Zapier so much as a different universe, and it is here for a specific reason: a hosted automation platform cannot reach a Modbus register, and no amount of paid plan changes that. For manufacturing, building management or a lab, this is the whole category rather than one option within it.
On the custody question it goes further than anything else here. There is no vendor at all — nothing is processed by anyone, because there is nobody in the path.
More than 4,000 community nodes cover an enormous range of devices and services, with JavaScript function nodes for whatever they miss. The costs are equally plain: no built-in team collaboration or versioning, uneven community node quality, weaker SaaS integrations than Make or n8n, and flows that become hard to maintain without deliberate discipline.
What Node-RED does better than Zapier
- Reaches hardware and industrial protocols — MQTT, Modbus — which no hosted platform can touch
- Runs on a Raspberry Pi or an industrial gateway, entirely offline if you want
- No vendor in the path at all, so there is nobody holding a credential to anything
- Apache 2.0 and free, with more than 4,000 community nodes
- JavaScript function nodes handle anything the node catalogue does not cover
Where Node-RED is a step down from Zapier
- No built-in team collaboration or versioning, which bites as soon as two people maintain flows
- Community node quality is uneven and unsupported
- SaaS integrations are much weaker than Make's or n8n's, because that was never the point
- Nothing here resembles Zapier's ten-minute path from idea to working automation
Standout against Zapier. It is the only tool in this comparison that can automate a machine as readily as an API, which is a job Zapier cannot be configured into doing.
- Which law reaches it. EU (Germany). Zapier is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Germany.
- Source code. Closed source, as Zapier is.
- Independently checked. GDPR, German DPA.
Best for: Mid-market companies whose integrations are ERP, SFTP and databases
Locoia GmbH is in Berlin with German hosting, a German processing agreement and German-language support, and its connector catalogue tells you who it is for: SFTP, databases and business systems rather than social scheduling and form builders. That is the mid-market integration problem — an ERP talking to a warehouse system — which is a different thing from marketing automation with the same shape.
Against Zapier the relevant difference is what happens when these workflows fail. A missed social post is an annoyance; a failed nightly order export is a business problem discovered by a customer, and Locoia's error handling and alerting are designed on that assumption. There are more than 250 connectors, a low-code builder with custom code steps, and an embedded option for software vendors shipping integrations inside their own product.
Self-hosting is available on request for enterprise customers, which softens the custody question without removing it for everyone. Pricing starts around €200 a month, which is a deliberate mid-market position rather than an oversight: too expensive for a small team, and the connector catalogue is a fraction of Make's or Zapier's with correspondingly thin community documentation.
What Locoia does better than Zapier
- German hosting with a German processing agreement and support in German
- Connectors aimed at the systems a mid-market business actually runs, including SFTP and databases
- Error handling and alerting built for workflows whose silent failure is a business problem
- Self-hosting available on request for enterprise, which Make does not offer at all
- An embedded option for software vendors shipping integrations in their own product
Where Locoia is a step down from Zapier
- From about €200 a month, far above Zapier's entry plans and wrong for a small team
- Around 250 connectors against Zapier's 9,000-plus
- Thin community and documentation, so you are relying on the vendor rather than on the internet
- Strongest in German-speaking markets, which shapes both the support and the catalogue
Standout against Zapier. It is the only platform here whose catalogue assumes your integration problem is an overnight SFTP drop rather than a social post.
- Which law reaches it. EU (Germany). Zapier is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Your own servers.
- Source code. Open source, where Zapier is not: you can read what it does rather than take the description on trust.
Best for: Small businesses running six automations that matter
Automatisch is a Berlin open-source project licensed under AGPLv3, and it is deliberately the smallest tool on this page. A visual flow builder, webhooks, scheduled triggers and multi-user support, running on a modest European virtual server. That is enough to replace what most small businesses genuinely automate, and nothing more.
Against Zapier the appeal is proportionality. n8n is the better engine and it is also a server, a queue and an upgrade schedule; for a company with six automations that is a disproportionate answer to a small problem. Self-hosted Automatisch makes you the sole data controller with no vendor holding a single credential, at roughly the cost of the smallest virtual machine your hosting provider sells.
The licence is worth noticing next to n8n's. AGPLv3 is open source in the conventional sense, where the Sustainable Use Licence is source-available with restrictions on offering the software as a service.
The constraints follow from the simplicity and are stated plainly: a small connector catalogue, a young project with a small team, and no code steps at all, so a transformation the builder cannot express has nowhere to go. The moment that ceiling is reached, the upgrade path is n8n. A hosted cloud exists from about €20 a month.
What Automatisch does better than Zapier
- Genuinely AGPLv3 open source, where n8n is source-available under a use-restricted licence
- Runs on a small European virtual server, so self-hosting is proportionate to a small business
- You remain the sole data controller, with no vendor holding any connection
- Simple enough for a non-technical user, which most self-hosted options are not
- Free self-hosted, with a hosted cloud from about €20 a month for teams that will not run a server
Where Automatisch is a step down from Zapier
- A small connector catalogue, so checking your own app list first is mandatory rather than advisable
- No code steps at all, so anything the builder cannot express has nowhere to go
- A young project with a small team, which is a real consideration over five years
- Will be outgrown quickly by anyone with more than a handful of workflows
Standout against Zapier. It is the only option here that is both a real open-source licence and simple enough for the person who built the Zaps in the first place.
- Which law reaches it. EU (Germany). Zapier is run from the United States, so the CLOUD Act obliges the provider to hand over data on a valid order regardless of which country the servers are in.
- Where the data sits. Germany/EU, or your own infrastructure.
- Source code. Closed source, as Zapier is.
- Independently checked. Open-source connector SDK.
Best for: Vendors whose customers ask them to connect to a CRM they have never heard of
elastic.io GmbH operates from Bonn as part of the Kyocera group and is not bought by an operations team at all. It is bought by a software company whose customers need to connect their CRM, their accounting package and their warehouse system to its product — and which would otherwise be building and maintaining every one of those connectors in-house, forever.
It reaches this page because a certain kind of Zapier account is a software vendor doing exactly that the hard way: a Zapier integration published as the official answer to "does it connect to our systems", with the customer paying for their own Zapier plan and the vendor with no visibility when it breaks. elastic.io is white-label by design, so the integration experience appears inside your product under your brand, with per-customer isolation as architecture rather than as configuration.
The deployment options are what make it viable where a third-party cloud is not acceptable: hosting in Germany and the EU, or on-premise and private cloud on your own infrastructure, which matters for a vendor selling into regulated industries. The connector SDK is open source even though the platform is proprietary, so connectors you write remain yours. Pricing is quoted by deployment and volume with an enterprise sales cycle, and the catalogue is smaller than the global iPaaS vendors.
What elastic.io does better than Zapier
- Fully white-labelled inside your own product, rather than telling customers to buy a Zapier plan
- On-premise and private cloud deployment, so a regulated customer's integration stays in their estate
- Per-customer isolation built into the architecture instead of assembled per account
- Open-source connector SDK, so the connectors you write belong to you
- German company and German hosting, as part of the Kyocera group
Where elastic.io is a step down from Zapier
- Irrelevant unless you sell software: it does nothing for automating your own back office
- Quoted pricing with an enterprise sales cycle rather than a plan you can sign up for
- A smaller connector catalogue than the global iPaaS vendors
- The platform itself is proprietary, with only the SDK published
Standout against Zapier. It is the only platform here that a software vendor can ship into a customer's own data centre under its own brand.
What actually breaks when you switch
Re-authorising is the work, and it cannot be delegated. Every trigger and every action needs a fresh grant in the new platform, performed by somebody with admin rights in the source system. Block an afternoon, do them in one sitting with the right people in the room, and use the list as an access review — you will find connections to systems nobody remembers approving.
Check the catalogue before anything else. Zapier reaches more than 9,000 applications and the alternatives reach hundreds or a couple of thousand. The application that breaks a migration is never the obvious one; it is the regional payroll system or the niche booking tool that only Zapier ever supported. Where the connector is missing, the honest options are an HTTP request you maintain yourself or leaving that one workflow behind.
And decide who owns the server before you self-host, by name rather than by team. The credential argument only holds if the machine is patched, backed up and watched; a self-hosted automation engine that nobody maintains is a worse outcome than the arrangement you left, because it holds the same keys with none of the operational attention.
Is the real risk the data or the access?
The access, and it is worth being precise because the two get conflated. The payloads passing through an automation platform are usually unremarkable — a name, an order number, a status change — and they are retained for a limited period in run history.
The standing authorisations are the durable thing. A Zapier account that has been running for three years holds tokens for every system anyone has ever connected, including the ones behind workflows that were switched off and never unlinked. That is a broader permission than most organisations have granted to any single supplier, and it was assembled one convenient afternoon at a time by people who each authorised exactly one thing.
The practical consequence is that the useful question is not "where is the data stored" but "who can act as us, in which systems, and who reviews that". Nothing about that question improves by moving to a European provider with the same architecture; it improves when the engine runs somewhere you control.
Is self-hosting realistic for a small team?
For a handful of workflows, yes, and the smallest option is the right starting point. Automatisch runs on a modest European virtual server, is AGPLv3 licensed, and does webhooks, scheduled triggers and a visual builder — enough to replace the six automations most small businesses actually run.
n8n is the serious version and is a real commitment: a server, a queue, upgrades, and somebody who notices when a workflow stops at two in the morning. In exchange there is no task meter, so a workflow running every minute costs exactly what a daily one costs, and code steps run without a runtime ceiling.
The honest arithmetic counts the person. A self-hosted server nobody patches is worse than a hosted platform that gets patched on a Tuesday, and if no one in the organisation wants that job, Make or Locoia is the better answer even though it keeps the credential question open.
Will the automations survive the move?
The logic will; the connections will not. Every trigger and action has to be re-authorised against the new platform, which means somebody with admin rights in each connected system sitting down for an afternoon and doing them one by one. There is no export that carries an OAuth grant, and there should not be.
Expect the catalogue to be the constraint rather than the builder. Zapier connects to more than 9,000 applications; n8n lists over 400, Make over 2,000, and Automatisch and Locoia far fewer. The tool you are certain of is not the problem — it is the fourth-ranked thing in your Zap list, the niche scheduling app or regional accounting package that only Zapier ever bothered to support.
Do the inventory before you commit. List every application that appears in any Zap, check each against the new platform's catalogue, and price the gaps as either an HTTP request you build yourself or a reason to stay.
Does moving to Make actually solve anything?
It solves two things and leaves the main one untouched, and this page would rather say so than sell you a flag. Make is in Prague, part of Celonis, and you choose an EU or US data centre at signup — so European processing and a European supplier are both genuinely available.
It also expresses logic Zapier cannot. Routers branch, iterators walk through arrays item by item, aggregators collapse the results back, and error handlers decide what happens when a step fails. A Zapier account fakes those shapes with duplicated Zaps that drift apart within a quarter.
What it does not change is custody. Make holds your credentials exactly as Zapier does, there is no self-hosting, and pricing is per operation so a branched scenario still consumes several per run. If your objection is the operations meter and Zapier's missing branching, Make is an upgrade. If your objection is that an external company holds standing access to everything, Make is the same arrangement in a friendlier jurisdiction.
Which one to pick
If the reason you are here is that one company holds standing access to every system you run, n8n is the answer and self-hosting it is the whole answer — the tokens sit on your hardware, there is no task meter, and code steps run for as long as your server allows.
If nobody is going to run a server and you want branching Zapier cannot do, Make is the honest choice, and this page will say plainly what it does not fix: the vendor still holds your connections, and the EU data centre changes which vendor rather than what they hold.
If your six automations are the whole requirement, Automatisch on a small European virtual machine is proportionate, genuinely open source, and cheaper than any plan on this page.
And be honest about the three entries that are not Zapier replacements. If your Zaps are scheduled pipelines, you want Kestra. If they touch hardware, you want Node-RED. If you are a software vendor shipping integrations to customers, you want elastic.io — and none of those problems was ever going to be solved by a different automation subscription.
Frequently Asked Questions
n8n if you want the engine on your own servers with code steps that are not time-limited. Make if you want a hosted platform with real branching and an EU data centre. Automatisch if six simple automations and a genuinely open licence is the whole requirement. Locoia if your integrations are ERP, SFTP and databases rather than marketing apps. Kestra, Node-RED and elastic.io are on this page for different jobs entirely.
The free plan covers 100 tasks a month with two-step workflows only. Professional runs from $19.99 a month billed annually up to $3,389 at the top of the range, covering 750 tasks to two million; monthly billing is $29.99 to $5,099, with annual described as 33% off the monthly rate. Team plans start at 2,000 tasks. Every successful action counts as a task, so complexity costs money on every run.
No European region appears on Zapier's pricing or plan pages, and its data processing addendum points at standard contractual clauses for transfers. Zapier, Inc. operates from San Francisco. That is a lawful and common arrangement; it is simply a different answer from a platform you run on your own infrastructure, or from Make, where an EU data centre is selected at signup.
Not quite, and the distinction matters if you plan to build on it. n8n is source-available under the Sustainable Use Licence, which permits free internal use while restricting offering n8n as a service to others. Automatisch is AGPLv3, Kestra is Apache 2.0 and Node-RED is Apache 2.0, all of which are open-source licences in the conventional sense. For internal automation the difference is academic; for a product, read the licence first.
No. There is no export format any of these import, and even if there were, the authorisations could not travel with it — every connection must be re-granted in the destination platform by someone with admin rights in the source system. Rebuild the workflows that matter, which is usually a third of the list, and use the exercise to delete the ones that have been failing silently.
The catalogue, first and mostly: more than 9,000 applications against n8n's 400-plus and Make's 2,000-plus, and the long tail is where the difference lives. Second, the ubiquity — anyone you hire has used it and every SaaS product publishes a Zapier integration before it publishes any other. Third, the speed of building something that works in ten minutes without thinking about infrastructure.
n8n, Automatisch, Kestra and Node-RED, all properly. Locoia offers self-hosting on request for enterprise customers and elastic.io supports on-premise and private cloud deployment. Make is cloud only. Running it yourself is the only configuration in which the credentials for every connected system sit on hardware you control, which is the whole argument of this page.
Yes, based in Prague and now part of Celonis, with an EU or US data centre chosen at signup. Choosing EU keeps both the processing and the stored credentials inside European jurisdiction. What it does not offer is self-hosting, so the vendor holds the connections either way — European ownership changes who that vendor is, not what they hold.
Stop being charged per action. Zapier counts every successful step, which means the filters and branches that make a workflow reliable also make it more expensive, and the incentive points the wrong way. A self-hosted engine charges nothing per run at all, so a workflow that fires every minute costs the same as one that fires daily, and you are free to build it properly.
Who worked on this review
Explore More European Alternatives
Discover privacy-focused European alternatives to other popular US tech services.