Security & Privacy

Zerodoc is built so the privacy promise is enforced by the architecture, not just stated in a policy. Documents are processed in memory and discarded the moment we respond — they are never written to disk and never stored.

Your document never touches a disk.

It’s processed in memory and gone before you finish reading the response — we can’t leak, lose, or be compelled to hand over what we never store. Zero-retention and EU processing are included on every plan, the free one included; they are never a paid upgrade.

Zero-retention by default

Every document you send to /v1/extract is held only in RAM for the duration of the request, then discarded. There is no document store, no archive, and no “retain for N hours” setting — by design, there is no code path that persists a document or its extracted content.

Each response includes an explicit "retention": { "stored": false } block as a machine-readable confirmation.

How processing works

  • Documents sent to the API go directly to our stateless EU processing service — the API hostname is deliberately not CDN-proxied, so document bytes take the shortest possible path to the machine that processes them.
  • Bytes live in an in-memory buffer that is overwritten on completion (success or error).
  • Uploads are size-capped and never spooled to a temporary file on disk.
  • OCR and field extraction run entirely on our own servers — a rules-first pipeline with local AI models as fallback. There is no third-party AI in the path and your documents never traverse an external service.
  • Usage telemetry records metadata only (page counts, size, timing) — never content.

Data residency & jurisdiction

Document processing runs on infrastructure in the EU — on every plan, including the free tier, never as a paid add-on. Zerodoc is a Rocsoft Ltd product based in the Isle of Man — an EU-adequate jurisdiction with a strong privacy reputation. For your documents we act as a data processor; we are the controller only for the account data described in our Privacy Policy.

What we store (metadata only)

The only data we retain is what’s needed to run accounts and billing: your email address, a one-way hash of your API key (never the key itself), and usage metadata (page counts, timestamps). We never store your documents or the data extracted from them.

API keys & transport

  • API keys are shown once at creation and stored only as a SHA-256 hash.
  • All traffic is encrypted in transit (TLS).
  • Per-key rate limits and monthly quotas protect against abuse.

Sub-processors

We use a small set of vetted providers:

  • netcup (EU, Germany) — hosts our stateless document processing; documents exist in memory only and are never stored.
  • Cloudflare — authentication, billing and usage metadata (EU-localized). Never receives documents sent to the API; test uploads made from the dashboard transit our Cloudflare-based control plane in memory only and are never stored.
  • Stripe — subscription payments; never receives your documents.
  • Resend — transactional (sign-in) emails; never receives your documents.

Compliance

Zerodoc is built to be GDPR-aligned. Our Data Processing Agreement is part of our Terms of Service and applies automatically to every customer. A SOC 2 Type II programme is on our roadmap. See our Privacy Policy for the full detail.

Reporting a vulnerability

Found a security issue? Email security@zerodoc.io — we welcome responsible disclosure.