Medical Documents in ONYRI Sanitize: What Gets Masked, and What You Mask Yourself
ONYRI Sanitize masks patient names, dates, contacts and ICD-10 codes in a medical document on its own, but not a stamp, a scan or a spelled-out diagnosis.
A medical letter, a lab report, or a hospital bill carries some of the most sensitive data a page can hold. ONYRI Sanitize can mask patient names, dates, contact details, national ID numbers, and ICD-10 diagnosis codes on its own. It cannot read a diagnosis spelled out in plain words, and it cannot see a stamp, a signature, or an X-ray image.
What Does ONYRI Sanitize Detect in a Medical Document?
Detection and masking run inside your browser. Nothing about the file, not its name, its text, or a value it finds, ever reaches ONYRI's server. About forty detectors scan the page across six families. For a medical document, the personal-data family does most of the work, as the table below shows.
| What's on the page | Example | Family in ONYRI |
|---|---|---|
| Patient name | Jane Example | Personal data |
| Date of birth or visit | 14 March 1987 | Personal data |
| Address, phone, email | 12 Example Street | Personal data |
| National social security number | French NIR, German or Austrian Sozialversicherungsnummer, Swiss AHV, US SSN | Personal data |
| ICD-10 diagnosis code | F32.1 | Personal data |
| Patient or file reference number | Patient ID: 48213 | Personal data |
Are ICD-10 Diagnosis Codes Masked by Default?
Yes. ICD-10 codes sit in the Default profile next to company names, so they are on from your first document, on Free and on Pro alike. ONYRI catches a code written with a decimal, such as F32.1 or Z73.0, wherever it sits on the page. A short code without a decimal, like J45, is only masked when a label nearby says ICD-10, CIM-10, or diagnosis. One letter and two digits alone could just as easily be a product reference.
The detector reads codes, not words. A line that says major depressive disorder or type 2 diabetes, with no code attached, passes through untouched. So does a drug name such as metformin. Skimming the detected-data list will not catch these. Search the document yourself, and mask the words by hand.
What Does ONYRI Sanitize Not Catch in a Medical Document?
- A diagnosis or a drug name written out in words, not a code
- A doctor's stamp or a clinic's letterhead logo
- A handwritten note or a signature
- An X-ray, a scan, or a photo placed on the page
- A barcode or QR code on a lab label
Every one of these needs a human eye, so ONYRI Sanitize added a manual masking tool for every plan, including Free, in late September 2026. A toolbar above the page offers three tools. Select text handles a single word or phrase. Search the document finds every instance of a term at once. Draw an area covers anything that is not text, such as a stamp or an image.
- Open the Masking tool bar above the page preview.
- Type the diagnosis or drug name into Search the document, then click Mask all 3 results.
- Or highlight the words yourself with Select text. Then choose Mask all 3 occurrences to cover every mention, or Only this one for a single spot.
- Choose Draw an area and drag a rectangle over a stamp, a signature, or an embedded X-ray.
- Changed your mind? Click a mask for Don't mask, or press Undo (Ctrl or Cmd+Z) for the last change.
Black Marker or Token Mode for a Medical File?
Black marker paints an opaque bar over each value and works on every plan. It fits a single letter you send once. Token mode replaces each value with a label instead, such as [NAME1] or [SECU1], and reuses that label every time the value repeats. It is Pro only, and it is pseudonymization, not anonymization: with the mapping, the person is identifiable again. Pick it when one patient's name must match across every page of the same document. There is no restore feature and no key file, so keep your original.
How Do You Review a Masked Medical Document Before You Share It?
Take a discharge letter for Jane Example, seen on 14 March 2026 for a fracture coded S52.5 in the file. ONYRI's page-by-page preview shows the masked result, and the Detected data list groups every hit by family, with one checkbox per value. Uncheck a value, such as the hospital's own address, to leave it visible. A warning appears if a value could not be placed on the page. Read it: it usually means an unusual line break kept a detected value from lining up with a box. Check the page and mask it yourself.
A medical document needs more care than most files you handle. ONYRI Sanitize catches the codes, the numbers, and the contact details on its own. It also gives you a manual tool for the stamp, the scan, or the words a machine cannot read. The review step is still yours: read the preview, check the warnings, and only then send the file.
Frequently asked questions
Does ONYRI Sanitize mask a diagnosis written in plain words?
No. The ICD-10 detector matches a code, such as F32.1, not a word like depression or diabetes. Type the exact word into Search the document, then click Mask all results to cover every mention.
Is a short ICD-10 code like J45 always masked?
Only near a label. A short code needs ICD-10, CIM-10, or diagnosis written close by, because one letter and two digits alone could also be a product reference. A code with a decimal, like J45.0, is masked wherever it appears.
Can Token mode restore the original values later?
No. There is no restore feature and no key file. Token mode keeps the same label for the same value across a document. That helps with a set of records about one patient. But the mapping only lives in your browser tab, and it disappears when you reload.
Does masking make a medical document GDPR compliant?
No. Masking reduces what a reader can see, but health data stays special-category data under Article 9 of the GDPR. Whether a specific use meets the rules depends on your case, so ask your data protection officer for anything beyond sharing one file.
Can ONYRI Sanitize mask a doctor's stamp or an X-ray image?
Not automatically. Use Draw an area to drag a rectangle over a stamp, a signature, or an embedded image. It works like text masking and lands in the same flattened export.
Sources & references
- GDPR Article 9 - special categories of personal datagdpr-info.eu
- What is special category data?Information Commissioner's Office (ICO)
- ICD-10 online classification browserWorld Health Organization (WHO)