What do the < (filler) characters in an MRZ mean?

The < character you see scattered throughout an MRZ is a filler character, and it serves a purely structural purpose: it pads out fields to their fixed, required length. Because the MRZ format is strict about how many characters each field occupies, a name field of a certain length, a document number of a certain length, and so on, any field that has fewer actual characters than its allotted space gets padded with < symbols to fill the gap.

You'll notice this most obviously in the name field, where << (a double filler) separates the surname from given names, and additional single < characters fill out any remaining space in that field after the full name is written. So "SMITH<<JOHN<<<<<<<<<<<<<<<<" is a way of encoding "Smith, John" while keeping the name field at its required fixed width, with unused character positions filled by the filler symbol rather than left blank.

The reason this matters, beyond simple curiosity, is that the fixed-width design is exactly what makes MRZ reading so reliable. Because every field always occupies the same number of character positions regardless of how long the actual name or number is, a parser doesn't need to guess where one field ends and the next begins.

It can rely entirely on position. The filler character is what makes that fixed-width structure possible even when the real data is shorter than the maximum allowed length.

For document scanning software, correctly interpreting filler characters means correctly stripping them out during parsing, so a name field doesn't come back to your system as "SMITH<<JOHN<<<<<<<<<<<<<<<<" but simply as "Smith John," properly formatted for whatever downstream system will use it, whether that's a KYC database, a hotel reservation system, or an age-verification check.

ScanDoc's MRZ parser handles this filler-stripping automatically as part of extraction, converting the raw MRZ string into clean, structured fields, name, document number, date of birth, and so on, with the ICAO formatting artifacts like < characters already removed, so the data arrives ready to use rather than needing additional cleanup on your end.