Check Whether Your PDF Really Is PDF/A

PDF/A is the ISO standard for documents that must stay readable decades from now, which it achieves by banning anything whose meaning depends on the outside world: fonts that are not embedded, colours with no profile to interpret them, encryption, external media. A file that opens perfectly today can still fail PDF/A, because "it looks right on my machine" is exactly the assumption the standard exists to remove.

This validator answers the question properly. Upload a PDF and it is checked against the standard with the same engine used throughout Documentize, and every non-conformance comes back listed rather than summarised: what is wrong, whether it is an error or a warning, which clause of the specification it breaks, and which internal PDF object and page it sits on. That last part is what makes a report actionable — "font not embedded" is a fact, "font 'ArialMT' is not embedded, clause 6.3.4, object 36, page 1" is something you can go and fix.

By default the file is judged against the standard it declares for itself. A PDF that claims PDF/A-2B is checked against PDF/A-2B, not against a level someone had to guess at, so a passing verdict means the file keeps the promise it makes. You can also name a level explicitly when compliance is a requirement handed to you rather than a property of the file — a regulator, a records system or a client asking for PDF/A-2B specifically. PDF/A-1 through PDF/A-4 are all supported, along with PDF/UA-1 for accessibility.

Identical problems are grouped by default. A single unembedded font is reported by the underlying engine once per use, which can turn one defect into dozens of rows; grouping collapses those into one entry with an occurrence count, so the list reflects how many things are wrong rather than how often they were noticed. The full report also downloads as JSON, ready to feed into a build check or an ingestion pipeline.