Security

How we keep
your data safe.

The same honesty policy as everything else on this site: what's genuinely in place, what relies on infrastructure we don't control, and what isn't built yet.

Last updated June 2026 · This page is rewritten whenever our architecture changes

The simplest security measure: nothing to steal.

Most security statements describe how a company protects its database. We don't have one. Every tool in the GDPRGard suite is a stateless page with no accounts and no server-side storage — so there's no central store of customer data that could be breached, because it doesn't exist.

That's not a workaround for not having proper security — it's a deliberate minimisation strategy. The fewer places data sits, the fewer places it can leak from.

🗄️
No database
Nothing you upload or generate is written to persistent storage on our side.
👤
No accounts
No passwords, no login sessions, no user profile data to protect or lose.
🔒
HTTPS everywhere
Every page and every API call is encrypted in transit, enforced by our hosting provider.
🔑
API keys stay server-side
Our Anthropic API key lives only in a serverless function, never in code your browser can read.

What we can actually back up.

Specific, verifiable measures — not generic security-page language.

Encryption in transit. Every connection to gdprgard.eu and every call to our backend functions runs over TLS. Plain HTTP is not served.
No client-side secrets. The Anthropic API key used to power the tools is stored as an environment variable in our serverless functions and is never shipped in any file your browser downloads — it cannot be extracted by viewing page source.
Origin checking on backend functions. Our serverless functions reject requests that don't originate from gdprgard.eu, reducing casual abuse of the API from other sites.
Data minimisation by default. Documents and text you submit to any tool are processed and discarded — not logged, not retained, not used for anything beyond generating your result.
Infrastructure-level protections from our hosting provider. Netlify provides DDoS mitigation and edge-network protections as part of its platform — we inherit these, we didn't build them ourselves.

Security we rely on, but don't control.

Three external services touch data as part of how the tools work. We chose each for their own track record, but we don't audit their infrastructure ourselves — for their specific security practices, their own documentation is the authoritative source.

ServiceSecurity info
Anthropic API Processes the content of what you submit. See Anthropic's Trust Center for their security and compliance posture.
Netlify Hosts our site and serverless functions. See Netlify's Trust Center.
Formspree Receives contact form submissions when you explicitly send one. See Formspree's security page.

What we haven't done yet.

Same principle as everywhere else on this site: we'd rather tell you what's missing than let you assume we have certifications we don't.

No independent security audit or penetration test has been performed on this codebase. It's been reviewed during development, but not by a third party.
No SOC 2, ISO 27001, or similar certification. These are formal, audited frameworks we have not pursued at this stage.
No bug bounty program. If you find a security issue, the best path today is reporting it directly by email (below) rather than through a formal program with set payouts.
No published incident response SLA. If something goes wrong, we will respond and disclose — but there's no contractual response-time guarantee in place yet.
No formal business continuity / disaster recovery plan. Given the stateless architecture, there's relatively little to recover — but this hasn't been documented as a formal plan.

None of this means the tools are unsafe to use for their intended purpose — quick checks and drafts for SMB compliance tasks. It does mean that if your organisation's procurement process requires formal certifications, we don't have them yet. Tell us if you need one; that's exactly the kind of demand that tells us it's time to build it.

Found a security issue?

We don't have a bug bounty program, but we take reports seriously and will respond. Please include "security" in your subject line so it doesn't get lost in general enquiries.

Report a security issue