Question 01
What is MRZ (Machine Readable Zone)?
The Machine Readable Zone, or MRZ, is the strip of text near the bottom of a passport data page or on the back of many national ID cards. It looks like a run of capital letters, numbers, and < filler characters arranged in fixed rows.
Unlike the printed name, photo, and address on the rest of the document, the MRZ was designed from the start to be read by machines rather than people. Each character sits in an exact position, so software doesn't need to guess where the date of birth ends and the document number begins.
It already knows, because every document that follows the standard uses the same layout.
The format was created by the International Civil Aviation Organization (ICAO) and is documented in ICAO Doc 9303, the same specification that governs biometric passports and electronic ID chips. Before this standard existed, border agents and clerks had to read and retype identity data by hand, which was slow and produced a steady stream of typos.
The MRZ solved that by giving every compliant document a predictable, scannable data block.
A typical MRZ encodes the document holder's name, nationality, document number, date of birth, sex, and expiry date, along with check digits that let a reader verify the data hasn't been altered. Passports use a two-line format called TD3, while many ID cards use a three-line format called TD1.
The font is OCR-B, chosen specifically because its characters are easy for scanners to tell apart even at low resolution.
For a business, the MRZ is what makes automated identity checks possible in the first place. A camera or scanner captures the zone, OCR software extracts the characters, and the result is a structured set of fields, ready to drop straight into a booking system, a bank's onboarding flow, or a hotel's front desk software.
ScanDoc's document scanning solution reads the MRZ this way as one part of a broader check. It cross-references the result against the visual zone and other data sources, so a single mismatched character gets flagged instead of silently accepted.
If you're building or buying an identity verification process, understanding the MRZ is the starting point. Most of what makes automated ID scanning fast and reliable traces back to this one standardized strip of text.
Talk to a document scanning specialist
Have a specific integration question, or want to see how this fits your onboarding flow? The ScanDoc team is happy to help.