What is document verification?

Document verification is the process of checking that an identity document is genuine, unaltered, and belongs to the person presenting it. A document verification system is the software that performs those checks at scale — reading the document, testing its internal consistency, and comparing what it finds against what the person has claimed.

It is one component of identity verification, not a substitute for it. A perfectly genuine document presented by the wrong person still produces the wrong answer, which is why serious systems always pair document checks with a link to the presenter.

A capable document verification system runs several independent checks and scores them together:

• Data-page checks. The machine-readable zone's checksums, barcode consistency, and the format of every field — document numbers, dates, check digits. Tampering almost always breaks one of these first.
• Authenticity checks. Fonts, layout, microprint texture and security features such as holograms or changeable ink, inspected for the properties the issuing document is known to carry.
• Coherence checks. Do the portrait, the name and the dates agree across the document's own surfaces? Has anything been digitally edited — resaved images, mismatched compression, altered text?
• Presenter checks. A live face matched against the document portrait under a liveness challenge — the only layer that connects the document to a person in front of the camera.
• Authoritative cross-checks. Where the issuer offers an interface, the document's details can be checked directly against the source of truth rather than only against the image.

That last layer matters more than it sounds: it verifies against a record instead of against a picture — an approach covered as one of the methods in identity verification solutions compared.

Document verification was designed around capture, but the web settled on upload: a file the user picks from disk. That single decision hands the attacker control of the evidence. A generated document image can be uploaded like any other file, and no amount of image inspection can prove the pixels came from a camera pointed at a real document.

The stronger pattern is guided capture: the camera runs inside the session with edge detection and glare checks, so the evidence is produced live and cannot be selected from disk. Add device integrity attestation — the app proving it is genuine and the operating system untampered — and the virtual-camera route closes as well. The full attack-and-defence picture is in AI-generated fraud and fake IDs.

It cannot prove personhood on its own. Without a live presenter, a document check only proves an image is well-formed. It cannot guarantee current details — addresses and names change between renewals. And it cannot answer the question nobody asks: what happens to the scan after the check. Every retained copy is breach inventory, which is why the retention question belongs in any honest evaluation of a document verification system.

In this service, document verification plays its part once — during wallet setup, combined with active liveness at the Standard level — and the durable output is a verified credential, not a retained image. What services receive instead of documents is spelled out in who can see your data.