What is ICAO Doc 9303?

ICAO Doc 9303, titled "Machine Readable Travel Documents," is the technical specification that defines how machine-readable passports, visas, and ID documents should be built. It's published by the International Civil Aviation Organization and is the foundational document behind everything related to MRZ: the field layouts, the OCR-B font requirement, the check-digit calculations, and, in more recent editions, the specifications for electronic passports with embedded RFID chips.

The document is organized into several parts, covering different document types and technical layers. Earlier parts define the physical and visual specifications for machine-readable passports and the MRZ formats discussed elsewhere in this FAQ.

Later parts, added as e-passports became widespread, define how the embedded chip stores biometric data, how it's digitally signed to prevent tampering, and how a reader authenticates the chip's contents against the printed data. This is why modern document verification systems often check MRZ data, visual zone data, and chip data together.

Doc 9303 effectively defines all three as parts of one interconnected system.

Doc 9303 also specifies the check-digit algorithm used throughout the MRZ, a defined arithmetic calculation applied to fields like document number and date of birth. Any reader that knows the algorithm can independently recompute the expected check digit and compare it to what's printed, which is the basis for detecting alterations without needing to contact any external database.

For businesses, Doc 9303 matters less as a document to read personally and more as the reason certain claims about MRZ reliability hold up. It's why an ID scanning vendor can say, with confidence, that a passport from one country can be read with the same logic as a passport from another, and why check-digit mismatches are a meaningful fraud signal rather than a guess.

ScanDoc's MRZ and visual zone extraction is built around ICAO 9303 compliance, which is part of why it can process documents from many countries without needing country-specific configuration for the underlying MRZ logic. The standard itself does most of that work, leaving the software to handle document classification, OCR of the visual zone, and cross-validation on top of it.