Healthcare & life sciences Special category data

Clinicians use AI every day. So does your patient data.

Your teams already lean on AI to draft discharge letters, summarise notes, code referrals and clean up research data. The risk isn't the tool — it's what gets pasted into it: patient records, diagnoses, test results, genetic and mental-health data. That's special-category data under GDPR. Tracehold catches it on the device, in the browser, so PHI never leaves the machine, and maps your AI exposure to GDPR (Art. 9), NIS2, ISO 27001 and SOC 2, with audit-ready reports.

PHI never leaves the browser GDPR Art. 9 · NIS2 · EU AI Act Built & hosted in the EU
about to be pasted into AI

REC · ON-DEVICE · CAUGHT BEFORE IT LEAVES

A patient name with a diagnosis
caught as special-category health data
Block
A patient's name, DOB and record number
caught as personal data, even free-form
Redact
A research dataset row with identifiers
caught before it reaches a public model
Block
The prompt stays in the browser. Only metadata — type, severity, timestamp — reaches your security team.
On the device
Where detection runs

We see what's about to go to AI right where the clinician works — before it leaves the machine.

PHI stays put
What leaves the browser

By default the prompt never leaves the browser. We only learn the type of finding, how severe it was, and when.

1,600+
Detection patterns

Across any AI tool or site your teams use — known and unapproved alike — including the free-form patient data that has no fixed shape.

EU
Built and hosted in Europe

Built and hosted in the EU (France), with EU data residency for everything that reaches our infrastructure.

[ Shadow-AI exposure · observe ]

The fastest way for patient data to leave is a prompt box

Generative AI didn't wait for a clinical safety review to arrive on the ward and in the lab — it came through the browser. Helpful, fast, and one paste away from sending special-category health data to a third party you never had a lawful basis to share it with. The intent is almost always to save time. The exposure isn't worth it.

Patient records & clinical notes

A clinician pastes a full case note — name, history, diagnosis, medication — into an AI tool to draft a referral or discharge letter. Each one is special-category health data under GDPR leaving your control.

Test results & research data

Researchers and analysts paste lab results, imaging reports or rows from a study dataset to "explain this finding" or "tidy up this table" — identifiers and outcomes going straight into a public model's prompt.

Code, integrations & documents

Developers paste integration code with embedded credentials into AI assistants to debug an HL7 or FHIR feed. Teams drop confidential protocols, ethics submissions and incident reports in to summarise them.

Said plainly: traditional DLP watches email and file shares — it never sees the prompt box in a browser tab. That's the exact channel AI runs on. Tracehold was built for that channel: it inspects what's about to be sent to an AI tool, on the device, before it leaves — so PHI never has to leave to be checked.
[ Detection coverage · 1,600+ patterns ]

Built to recognise the data a health organisation can't afford to lose

With 1,600+ detection patterns running across any AI tool or site your teams use, Tracehold recognises both structured identifiers with a fixed shape and the free-form clinical detail that has none — and reads the text inside screenshots and scanned reports, on the device, by default.

Special-category health data

Diagnoses, conditions, medications, mental-health and genetic detail, and clinical narrative — recognised as the special-category data GDPR singles out for the strongest protection, even when it's written in free-form prose.

Patient identifiers

Names, dates of birth, addresses, national and patient record numbers, and contact details — including names and addresses that follow no fixed format and slip past pattern-only tools.

Credentials & confidential docs

API keys and tokens hidden in pasted integration code, plus internal protocols, ethics submissions and incident reports — caught so a quick "summarise this" doesn't quietly export them.

[ On-device pipeline · PHI stays put ]

PHI never has to leave the browser to be protected

In healthcare the usual DLP bargain — "send your data to our cloud so we can inspect it" — is exactly the thing you're trying to avoid. Tracehold flips it: the check happens on the device, in the browser, so the patient data being protected never leaves the clinician's machine to get checked in the first place.

Detection runs locally

The prompt your team types is inspected right in the browser. By default it never leaves the tab — not to an AI tool, and not to us.

Only metadata in managed deployments

When Tracehold is centrally managed, only metadata — the type of finding, its severity and a timestamp — reaches the backend. The prompt content itself does not leave the browser by default.

Free plan stays entirely on-device

The free browser extension runs with no account and no backend — everything happens on the device. Nothing about what your staff type is sent anywhere.

Files already on the device, not just what's typed

The desktop agent can also scan PDFs, Word documents and spreadsheets already saved on a device — a discharge letter or a research export — for the same sensitive detail, before any of it is dragged into a prompt box.

The honest version

On the desktop, an organisation may choose to keep the exact text behind an alert so an admin can tell a real leak from a false alarm. Where that option is enabled, the text is encrypted and can only be revealed through an audited, permission-controlled action — and that's configured by you, not by us.

No silent copies, no master back door. We'll always tell you exactly where a sensitive item can exist and how it's protected. See how we protect your data →

[ GDPR Art. 9 · NIS2 · EU AI Act ]

Uncontrolled AI use touches the rulebooks that matter most in health

Health data sits at the strictest end of data protection. A single paste can trip more than one obligation at once.

GDPR — Art. 9 special category

Health data is special-category data under Article 9, with a higher bar for processing and sharing. Pasting it into a public AI tool can be an unlawful transfer to a third party. Catching it on the device keeps it from ever leaving your control.

NIS2

Healthcare is named among the essential and important sectors under NIS2, with stronger cyber-risk management and accountability. Demonstrable control over how data leaves to AI tools — with an audit trail — is part of showing you manage the risk.

EU AI Act

As clinical and research teams adopt AI, you need to know where and how it's used, and govern it. Tracehold maps your actual AI use so you can govern it rather than guess at it.

[ Compliance mapping · live readout ]

Compliance mapping you can put in front of the board

Tracehold maps your real AI exposure to GDPR, NIS2, ISO 27001 and SOC 2 — with GDPR Article 9 special-category data front and centre for health, and the EU AI Act's governance expectations built in. Instead of an annual questionnaire, you get a clear, always-current view that updates as your risk does — and the evidence behind it when an auditor or regulator asks.

Risk, not theatre

The mapping reflects what your people are actually doing with AI today — what's being caught, how often, how severe — not a point-in-time checklist that's stale by lunchtime.

Evidence on demand

Every finding leaves an audit trail with just the essentials — type, severity, timestamp — so you can show diligence to a regulator or DPA without exposing the underlying patient data.

Aligned, and honest about it

Tracehold is aligned with and mapped to ENS, PCI-DSS and SOC 2, with ISO 27001 certification in progress. We tell you exactly what is certified and what is mapped — never more.

From "we think we're fine" to "here's the picture"

Most health organisations can't answer a simple board question: how much patient data is going to AI right now? Compliance mapping answers it — and shows the trend as your controls take effect.

The mapping covers GDPR, NIS2, ISO 27001 and SOC 2 in one view, with EU AI Act governance built in. See compliance in depth →

[ Deployment · MDM-pushed ]

Weeks, not months — no proxy, no network surgery

A hospital or trust can't bolt a new inline proxy into a clinical network on a whim. Tracehold rolls out as a browser extension you push with the device management you already run — so the security team stays in control and the network stays untouched.

Pushed via your MDM

Deploy the extension centrally through Microsoft Intune or any standard MDM, tied to your identity provider — Microsoft Entra ID or Okta. No agent your endpoint team has to babysit, no inline gateway.

Start observe-only

Begin in a pilot that watches and reports without blocking anyone. You see exactly what would be caught — and where your real exposure is — with zero disruption to clinical work before you turn on enforcement.

Plugs into your stack

Findings and alerts export to your SIEM, with real-time alerts to Slack or by email. Hosted in the EU (France), with EU data residency for everything that reaches our infrastructure.

Most teams go from first call to protected in 2–4 weeks. Detection runs on the device; only metadata reaches the backend by default.
[ Plans · published pricing ]

Start free, scale to governed

Honest, published pricing. Begin on the device for free, add managed visibility and compliance reporting when you're ready.

Free

Runs entirely on the device. No account, no backend — nothing your staff type is sent anywhere.

See what's included

Insight — €40/user·mo

Managed visibility into AI exposure across your teams, with metadata-only reporting and live compliance mapping.

See Insight

Governance — €60/user·mo

Full enforcement, audit trail and integrations for security and compliance teams that need to prove control.

Talk to our team
[ FAQ · buyer questions ]

What healthcare buyers ask first

Does patient data ever reach Tracehold?
No, not by default. Detection runs on the device, in the browser, and the prompt itself never leaves the tab — so PHI doesn't have to leave the machine to be checked. In managed deployments only metadata — the type of finding, its severity and a timestamp — reaches the backend, never the prompt content. The free extension goes further still: it runs entirely on the device with no account and no backend at all, so nothing about what your staff type is sent anywhere.
How does this fit GDPR Article 9 for health data?
Health data is special-category data under Article 9, with a higher bar for processing and sharing. Pasting it into a public AI tool can be an unlawful transfer. Tracehold recognises special-category health data — including free-form clinical narrative — and catches it on the device before it leaves, so it stays within your lawful basis. The compliance mapping reflects that exposure against GDPR, NIS2, ISO 27001 and SOC 2, with EU AI Act governance built in.
Will it slow clinicians down or break their tools?
No. There's no inline proxy and no network change — it's a browser extension running locally. You can start in observe-only mode so nothing is blocked while you measure exposure, then enable enforcement on the policies you choose, with clear in-context messages rather than silent failures.
Are you certified — ISO 27001, SOC 2?
We're honest about status. ISO 27001 certification is in progress — not yet held. Tracehold is aligned with and mapped to SOC 2, PCI-DSS and ENS, but is not certified against them — we never claim certification we don't hold. We're happy to walk through the detail with your information governance and risk teams.
Where is patient data actually hosted?
Tracehold is built and hosted in the EU (France), with EU data residency for anything that reaches our infrastructure. And detection itself runs on the device, so by default the prompt content never reaches us in the first place — only metadata does, in managed deployments. The free browser extension goes further: it runs entirely on the device with no account and no backend at all.
How fast can we be protected?
Weeks, not months. You push the extension through your existing MDM, no proxy or network changes required. Most teams go from first call to protected in 2–4 weeks, starting with an observe-only pilot. Pricing is published up front: Free, Insight (€40/user·mo) and Governance (€60/user·mo).

See what your organisation is already sending to AI

Start an observe-only pilot and get a clear, honest picture of where patient and research data is leaking to AI tools — with zero disruption to clinical work, and PHI that never leaves the browser. References available under NDA.