Privacy-first AI infrastructure.

← Use cases

AI help with the paperwork. Nothing about the patient leaves.

Draft the letter and tidy the summary with names, dates of birth and record numbers replaced on the device, never sent.

Start free in the browserDocuments and folders

What you have

Discharge summary for Tomas Keller, born 14/03/1961, NHS 943 476 5919, of 12 Rowan Court, Leeds.

What the assistant receives

Discharge summary for ⟦PERSON_1⟧, born ⟦DATE_1⟧, NHS ⟦NHS_NUMBER_1⟧, of ⟦ADDRESS_1⟧.

Synthetic. The reply comes back with the tokens in it and the real values are restored on your machine.

What gets masked
Names, dates of birth, addresses, contact details and identification numbers. In pseudonymised mode a shifted date stays a plausible date, so the note still reads as a note.
Where it runs
The browser extension does the whole round trip with no network access, which you can test by pulling the connection. A cloud de-identification API has to receive the identifiable text before it can remove anything.
Measured, not promised
Precision, recall and what still slips through are published per language, along with the honest part: no detector catches everything, so read the numbers before relying on them.

Clinical work generates exactly the text an assistant is good at, and exactly the text with the least room for error in handling. A patient note is identifying almost by definition.

The pasting already happens. The choice is not whether AI touches clinical text, but whether anything identifying is still in it when it does.

What detection catches, and what still slips through, is measured and published on the benchmarks page.

What people ask first

Is masking patient identifiers the same as de-identification under health privacy rules?

Rules like HIPAA and the GDPR define their own standards for de-identification and pseudonymisation, and meeting them is your organisation's determination to make. What Velum changes is the factual baseline: identifying details are replaced locally before any text reaches an AI provider, and the re-identification mapping never leaves your machine.

Does it recognise medical record numbers and health-specific identifiers?

Detection covers names, dates, addresses, contact details and a wide range of identification and account number formats, with country packs for Spain, the US, UK, France, Germany, Turkey and Italy. The benchmarks page lists measured detection quality per language, which is the reliable way to see what is and is not caught in your material.

Can I use this without sending anything to Velum?

Yes. There is no Velum account and no Velum server in the data path. The extension runs with zero network access. The desktop app makes a short, enumerated list of outbound calls, none of which carry document content, and shows that list in its own Protection tab.
Request a demo

See it work on your own data.

Tell us what you work with and we will walk you through it on a call, on your own files. Nothing leaves your machine while we do.

Your details mask themselves as you go. That is Velum, running in this page. Press the eye to unmask. We still receive them in full.

Your address is used to reply to you, and for nothing else.