Document verification
Read an identity document, confirm it is genuine, and return structured, checked data — in one API call. Zanyara handles passports, national ID cards, driving licences and residence permits, using every independent copy of the identity a real document carries so a decision rests on more than a single photo of a page.
What a real document gives you — and how Zanyara uses each part
- Visual inspection zone (VIZ): the printed fields are read by a trained document-OCR processor — not a general language model — so field extraction is deterministic and auditable.
- Machine-readable zone (MRZ): the encoded lines are parsed and their check digits validated, giving a machine copy of name, document number, dates and nationality to compare against the print.
- NFC / eMRTD chip: on ePassports and many eID cards, the chip’s signed data is read on the applicant’s phone and verified server-side (passive authentication).
- 2D barcode: the PDF417 barcode on some driving licences and ID cards is decoded as a further machine source.
Extract, corroborate, decide
- ClassifyThe document type and issuing country are detected automatically (you can pass a hint to tighten capture guidance).
- ExtractThe printed fields are OCR’d and normalised; the MRZ, chip and any barcode are decoded.
- CorroborateThe independent sources are cross-checked — the print, the MRZ, the chip and the barcode should agree, and their check digits and signatures should validate. Disagreement is a fraud signal, not a silent pass.
- DecideA structured result and a clear pass / refer / fail outcome are returned, with the reasons attached for your reviewers and audit trail.
Supported documents
- Passports — including ICAO ePassports with NFC chip reading and passive authentication (see NFC chip verification).
- National ID cards — including eID cards with chips and cards carrying 2D barcodes.
- Driving licences — including the UK driving licence.
- Residence permits — see residence-permit verification.
A document is one signal, not the whole answerVerifying the document proves it is genuine and unaltered. To bind it to the person in front of you, pair document verification with face match and liveness so the live selfie is compared against the document (or the chip’s signed photo).
Frequently asked questions
- Which identity documents can Zanyara verify?
- Passports, national identity cards, driving licences and residence permits from a wide range of countries. Where a document carries an NFC chip (ePassports, many eID cards) Zanyara reads and cryptographically verifies it; where it carries a 2D barcode (some driving licences and ID cards) Zanyara decodes that too.
- Is the data extracted by an LLM?
- No. Extraction runs on a trained document-OCR processor, not a general-purpose language model, so field reads are deterministic and auditable. The machine-readable zone, any chip and any barcode are then used as machine sources to corroborate the printed text rather than trusting a single read.
- How does Zanyara check a document is genuine, not just readable?
- It cross-checks the independent copies of the identity that a real document carries — the printed visual zone, the machine-readable zone, the chip and any barcode should all agree, and their check digits and (for chips) cryptographic signatures should validate. A forger has to keep every one of those in sync, which is far harder than editing the photo page.
- Do I have to tell Zanyara the document type and country up front?
- No — Zanyara classifies the document type and issuing country automatically. You can optionally pass a hint (for example when your flow only accepts UK passports) to make capture guidance stricter, but it is not required.