n8n self-hosted or cloud: what it costs and what it risks
03 November 2026
n8n is a workflow automation tool you can run in two ways. n8n Cloud is hosted by n8n for a subscription. Self-hosted means you run it on your own server, or have someone run it for you. The choice changes your costs, your data protection position and who gets the call when a workflow fails at 3 a.m.
We run self-hosted n8n in production for our own systems and for clients, so this comes from operating it. Check current prices on n8n’s site before you decide: they change, and we do not quote numbers here that will go stale.
How much does n8n cost?
n8n Cloud is a subscription in tiers. The unit that matters is the workflow execution: one full run of a workflow counts as one execution, however many steps it contains. Higher tiers include more executions per month and more active workflows.
Self-hosting the community edition has no licence fee. n8n publishes it under its own Sustainable Use License, which allows internal business use but is not a standard open-source licence. Features such as single sign-on, advanced permissions and some collaboration tools sit in paid plans, including for self-hosted installs.
Two points people miss:
- Execution-based pricing is kind to workflows with many steps and unkind to workflows that fire thousands of times a day, such as one run per incoming email.
- Build time is usually a bigger line than the subscription.
How much does self-hosted n8n cost?
The licence costs nothing, but running it does:
- A server. A small virtual machine in an EU data centre covers many teams. Heavy workloads need more memory and a proper database.
- A database and storage for workflow data and execution history, with a retention policy so it does not grow forever.
- Backups that you have actually restored at least once.
- Updates and patching. n8n releases often and publishes security advisories, like any software. Self-hosting makes you responsible for applying them.
- Monitoring and alerts, so a failed workflow reaches a person instead of sitting in a log.
For a team without anyone who has run a server, those tasks are the cost. For a team with an engineer or a partner who does it already, the marginal cost is small. Compare the subscription with the hours of operating it.
Is n8n GDPR compliant?
GDPR compliance depends on your set-up: where it runs, who processes the data and how long it is kept.
n8n Cloud is a processor under GDPR article 28. You need a data processing agreement, and you need to check where the instance runs and which sub-processors touch the data. Ask for both in writing and read them. For workflows that carry personal data, that is the minimum.
With self-hosting, you choose the location. Put the server in an EU data centre, keep credentials and execution data there, and the transfer question is smaller. You are still the controller, you still need your own processing records, and you still need a processor agreement with your hosting provider.
Two things apply to both options. First, n8n stores execution data, including the content that passed through each step. Set retention so personal data is not kept longer than you need it. Second, if a workflow sends data to an AI model provider, that provider is a processor too, and it needs its own agreement. Our GDPR checklist for AI automation covers that chain.
This is a technical summary, not legal advice. Your data protection officer or counsel signs off on your own set-up.
How secure is n8n?
Both options can be secure, and both can be set up badly.
On Cloud, n8n handles the infrastructure. You still control who has access, how credentials are scoped and what each workflow can reach.
Self-hosted, you own the whole surface. The basics we apply on every install:
- No public access to the editor. Put it behind a VPN or an authenticated proxy, and expose only the webhook endpoints that have to be public.
- Encryption key and credentials stored outside the workflow files, backed up separately.
- Least-privilege API keys, one per integration.
- Timely updates and a short log of what was changed and when.
- Error handling in every workflow: retries, and an alert on final failure.
The weakest point in most n8n installs is an editor exposed to the internet and an API key with more rights than it needs.
What is the difference between n8n and Make?
Make is cloud-only. You cannot host it yourself. n8n can be self-hosted, which is the main reason EU companies with data protection requirements look at it first.
The other differences are practical. Make has a polished visual builder and charges by operations. n8n charges by executions on Cloud, and lets you write code inside workflows when the visual nodes are not enough. That helps when you build real systems, with custom logic and AI steps, and costs you if nobody on the team can read code.
For simple, low-volume automations between popular apps, Make is often faster to start. For workflows that carry personal data, run at volume or contain AI agents, the ability to self-host and add code usually decides it.
When each option makes sense
Choose n8n Cloud if you have no one to operate a server, the workflows are modest in volume, the data is low-sensitivity, and you want to start this week.
Choose self-hosted if the workflows carry personal or confidential data, volume is high, or you need control of location and retention.
Choose a managed self-hosted set-up if you want the second without hiring for it. That is how we run it: a server in the EU, monitoring, retries, and an audit trail.
Our LinkedIn outreach engine sends from 26 points, all routed through one safety gate that guarantees each message is sent at most once and writes every attempt to an audit table. A guarantee like that is much easier to build when you own the database and the execution environment.
If you want n8n built, hosted and kept running, see our n8n agency service. If you are still deciding between tools, tell us your workflows in the free automation audit and we will say which option fits.