Contracting party
The provider is n8n GmbH, Novalisstr. 10, 10115 Berlin, Germany (HRB 212509 B).
n8n is seen as the European answer to Zapier and Make, and much of that holds: a Berlin GmbH, cloud in Frankfurt, self-hosting possible. The open questions are in the execution logs and the AI nodes.
The provider is n8n GmbH, Novalisstr. 10, 10115 Berlin, Germany (HRB 212509 B).
On hosted plans, data is stored in the EU, on servers in Frankfurt. Self-hosted, it is wherever you run n8n.
For transfers outside the EU, n8n names adequacy decisions such as the EU-US Data Privacy Framework where applicable, and Standard Contractual Clauses.
According to the sub-processor list (updated 22 September 2026), OpenAI and Anthropic as AI providers and the US affiliate n8n US Inc., among others, process data in the US.
The cloud's self-serve terms state that the data processing agreement applies.
The cloud prunes execution logs automatically: Start and Starter after 7 days (max 2,500 executions), Pro after 30 days (max 25,000), Enterprise with no time limit (max 50,000).
Self-hosted, n8n prunes executions by default after 336 hours (14 days) or beyond 10,000 saved executions. Annotated executions are never pruned.
Redacting execution data hides content in the interface and API. In the database it remains readable to anyone with direct access.
For self-hosted installations, n8n's documentation says it is neither controller nor processor, because it does not manage your data.
Self-hosted instances send usage data to n8n, including user identifiers and a truncated IP address, unless you opt out.
This page is not legal advice and we are not a law firm. We report what the vendor itself documents, with a date. Vendors change their terms; the linked source in its current version always prevails.
The cloud is in Frankfurt, but an AI node using OpenAI or Anthropic sends the data to the US. Every workflow with AI nodes therefore needs a look at which personal data flows into them.
When self-hosting, n8n is not your processor, but your server provider is. Add telemetry that runs by default and annotated executions that are never pruned.
Cloud or self-hosted. In the cloud, the data processing agreement with n8n applies; self-hosted, you need a server in the EU and an agreement with its provider.
Switch off saving successful executions or shorten retention, switch off telemetry when self-hosting, use annotations sparingly.
Which workflows send data to AI providers in the US, whether that is necessary, and how it is described in your records and privacy notice.
No tool is compliant on its own; it depends on the setup. n8n starts well: a provider in Berlin, cloud in Frankfurt, self-hosting possible. What remains open are execution logs with personal data and AI nodes that send data to the US.
According to the pricing page, in the EU, on servers in Frankfurt. Self-hosted, the data is wherever you run n8n.
No. n8n is then not your processor, but your server provider is. Telemetry runs by default, and every node that calls an external service sends data there.
In the cloud, 7 days, 30 days or, on Enterprise, no time limit, depending on plan. Self-hosted, 14 days by default, and annotated executions are never pruned.
The setup costs € 890 for your whole stack. The other tools we have checked the same way:
Tell us which tools you use and where your contacts come from. In the first call we will tell you whether € 890 covers it or whether more needs fixing, honestly either way.
Request the setup