Your Zap holds the passport scan for up to 69 days

Onto Amazon Web Services infrastructure in the United States — and it stays there for weeks after the Zap has finished running.

Zapier’s own data privacy page, read on 24 September 2026 and itself undated, states that its production infrastructure runs on AWS in the United States, and that there is no option to keep processing inside the EU. The same page sets out how long Zap content and history are kept: between 29 and 69 days. A shorter retention window of 7 to 30 days exists, but only on the Company and Enterprise plans.

Which means a Zap that lifts a passport scan out of an inbox and files it in a folder is not passing the file through. It is taking a copy. On most plans that copy outlives the task by more than a month, on infrastructure in another country, under a vendor most firms have never assessed.

Ten minutes with your Zap list

Open the dashboard and read every Zap you have. For each one, answer a single question: does anything in this payload contain an identity document, or the data printed on it? Passport and driving licence images are the obvious cases. The ones people miss are bank statements, proof of address, a form field holding a National Insurance number, and any trigger that grabs a whole email — attachments included — rather than named fields.

Most small firms find two or three. That short list is the one that deserves a decision rather than a habit.

The decision is rarely to abandon automation. It is usually to keep the document in one system and let the automation carry a reference to it. Ask whichever tool holds the file — ClientLoop, Content Snare, your existing document management system — whether it can hand an automation an ID and a short-lived link instead of the file itself. Running n8n on your own infrastructure is the other route, and it moves the hosting question onto you, which is rather the point.

Either way, the page to read first is the vendor’s data privacy page, not its integrations page.