Anyone looking to automate HR processes almost always runs into cloud tools first: Zapier, Make, Power Automate. For many HR use cases — onboarding checklists, application intake, deadline reminders — that's technically more than sufficient. The problem is rarely the technology, it's the question of where the data sits in the meantime.
n8n is an open-source automation platform that can run as a Docker container on your own server or your own cloud instance. That makes it one of the few options where personal data never has to leave your own area of responsibility.
What technically sets n8n apart from Zapier & co.
n8n is open source under a fair-code license (Sustainable Use License). In practical terms: the code is open, and a company can run the software on its own infrastructure instead of renting it as SaaS from a third party. n8n itself also offers a hosted cloud variant, but the self-hosting option is the real difference to Zapier or Make, which run exclusively as cloud services.
Technically, n8n is a visual workflow editor: nodes (triggers, actions, conditions) are connected by drag and drop. Over 400 prebuilt integrations exist, plus HTTP-request and code nodes for everything that doesn't have a ready-made integration — for example an internal HR system without an official API.
Why that matters for HR data
An onboarding workflow typically processes names, private contact details, ID data, sometimes health data (initial medical check) or salary information. If this workflow runs on a US cloud SaaS, questions about international data transfers, data processing agreements and — depending on the works agreement — a co-determination requirement under German law (§ 87 Abs. 1 Nr. 6 BetrVG, technical systems for monitoring behaviour or performance) come up.
On a self-hosted n8n running on your own IT infrastructure or with an EU hosting provider, the data stays within your own area of responsibility. That doesn't replace a data protection impact assessment or a data processing agreement with the hosting provider, but it reduces the number of involved third parties to a manageable list.
What self-hosting means in practice
Self-hosting n8n means: a server (your own IT or an EU cloud provider), a running Docker container, updates, backups, access rights. That's not a weekend project for a single person in HR — this needs the IT department or an external provider that takes it on and maintains it.
In practice, that means collaboration: HR brings process knowledge (which steps happen during onboarding, who needs to be informed when), IT brings hosting, access concept and maintenance. That split is also why an n8n project rarely stands within a few days, but rather runs over several weeks, in parallel with alignment with the works council and data protection.
Where n8n hits its limits
n8n doesn't replace core systems like an applicant tracking system or a personnel file — it connects them. And n8n is not a language model: anyone who wants to add AI functions (text summarisation, classification) connects a separate model for that, via a node. This is often exactly where it gets decided whether a cloud language model is an option or whether — for the same data protection reasons as the hosting itself — a locally run model becomes necessary. That decision belongs in the same alignment round as the hosting decision itself, not afterwards.
