How confidence is scored Each match is scored by what was found (Low 65 / Medium 75 / High 85), following the SIT's own rule pack — a credit card passing the checksum alone scores Low, and High only with a nearby keyword or expiry date. See the Reference guide.
Coverage (checksum-friendly subset of Standard SITs): — SITs — — with a reproduced checksum, — pattern + keyword. Use the Match against dropdown to see the full list.
- Reproduced checksums: Luhn (credit card, Canada SIN, South Africa ID), ABA weighted mod-10, IBAN mod-97, Verhoeff (India Aadhaar), mod-11 (Brazil CPF/CNPJ), mod-23 letter (Spain DNI), 11-test (Netherlands BSN), weighted mod-10 (Poland PESEL).
- No-checksum SITs (patterns auto-derived from Microsoft's documented formats) use the pattern plus a supporting-keyword check within Purview's ~300-char proximity window.
- Excluded — SITs whose checksum is proprietary and not reproducible from public docs (e.g. Australia Company Number, Australia Tax File Number, and many national IDs).
Approximate — not a substitute for Microsoft Purview. This tool covers the
checksum-friendly subset of Standard SITs: those with no checksum, and those whose
checksum is a well-known public algorithm we can reproduce (Luhn, ABA, IBAN mod-97, Verhoeff,
Brazil/Spain/Netherlands/Poland national-ID checks). SITs whose checksum is proprietary and
can't be reproduced are deliberately excluded to avoid misleading results. It applies
each SIT's regex, the checksum where reproducible, and a supporting-keyword check within
Purview's ~300-character proximity window. It does not reproduce the exact Boost.RegEx dialect,
strong date/format validation, or confidence scoring — results are indicative only and should
be confirmed in Purview.