Home / Learn
By PDFWix Editorial Team · October 2025
When PDF/A is required for long-term archiving, how it differs from regular PDF, which validators actually work, and the gotchas that fail compliance audits.
Regular PDFs are fragile over decades. They can reference fonts the reader doesn't have, embed JavaScript that future viewers won't execute, depend on encryption keys that get lost, or link to external content (web fonts, remote images) that disappears. Open a 1998 PDF in 2026 and you'll often see substituted fonts, missing graphics, or a broken interactive layer.
PDF/A (ISO 19005) is a self-contained subset of PDF that bans every feature liable to break later — no JavaScript, no external content, no encryption, no transparency in the older levels — and requires every font to be embedded in full. The result is a file that any conforming viewer, today or in 2070, will render the same way.
PDF/A comes in three generations (1, 2, 3) and three conformance flavours (a, b, u). The number is the PDF spec it's based on: PDF/A-1 = PDF 1.4, PDF/A-2 = PDF 1.7, PDF/A-3 = PDF 1.7 with embedded files allowed.
The letter is the strictness: -b (basic) guarantees visual reproducibility — the file will look the same forever. -u (unicode) adds the requirement that all text maps to Unicode, so the document is reliably searchable and copy-pasteable. -a (accessible) adds full PDF/UA-style tagging — logical structure, alt text, reading order — for screen-reader use.
Healthcare records under HIPAA retention rules, court filings via the US federal CM/ECF system and most EU e-justice portals, government records under NARA and EU member-state archive laws, banking records under SOX and Basel III retention, and academic theses at most universities — all increasingly mandate PDF/A-2 or -3 for any document held longer than seven years.
The free reference validator is veraPDF (verapdf.org) — open-source, audits against every PDF/A flavour, runs on Windows/Mac/Linux, and produces a per-rule failure report. Adobe Acrobat Pro's Preflight (Tools → Print Production → Preflight) is the paid alternative and slightly stricter on edge cases.
If a file fails, the most common culprits are unembedded fonts, transparency in PDF/A-1, missing XMP metadata, and external hyperlinks to web resources — all easy to fix in the source application before re-exporting.
Regular PDF allows everything the spec ever shipped: JavaScript, encryption, audio and video, external font references, transparency, layers, attached files and links to remote URLs. PDF/A bans nearly all of these because each one is a future failure waiting to happen — a JavaScript that won't run, a font that can't be found, an encryption key nobody can locate.
PDF/A also requires what regular PDF treats as optional: every font fully embedded (not just subsetted in the older flavours), XMP metadata describing the file, a device-independent colour space (typically ICC-tagged sRGB or CMYK), explicit text encoding mapped to Unicode, and a declared document language. A regular PDF that happens to satisfy all those requirements is essentially a PDF/A — you just need to add the conformance flag and validate.
US federal courts via CM/ECF, the European Court of Justice and most national courts in the EU require PDF/A for electronic filings. The US National Archives (NARA), the UK National Archives, the German Bundesarchiv and most national libraries mandate PDF/A-2 or PDF/A-3 for any document held longer than seven years.
Outside government, ISO 19005 is the de facto standard inside regulated industries: healthcare records under HIPAA retention, banking records under SOX and Basel III, pharma submissions to the FDA and EMA, and most university thesis repositories. If your industry has a multi-decade retention requirement, PDF/A is almost certainly already in your compliance manual.
Adobe Acrobat Pro: File → Save As Other → Archivable PDF (PDF/A). Free alternatives include LibreOffice (File → Export As → Export as PDF → tick 'PDF/A' and pick the level), Microsoft Word 2016+ (Save As PDF → Options → ISO 19005-1 compliant) and Ghostscript on the command line. After conversion, always validate with veraPDF — silent failures (claiming PDF/A without actually conforming) are common.
PDF/A-2u is the most common choice for general archiving — it guarantees visual reproducibility and reliable Unicode text search. Use PDF/A-2a or -3a when the document must be screen-reader accessible, and PDF/A-3 when you need to embed source files (XML invoices, original Word docs) inside the archive.
Use veraPDF (verapdf.org) — the open-source reference validator. It audits against every PDF/A flavour, runs on Windows, Mac and Linux, and produces a per-rule failure report so you can see exactly what's non-conformant.
Unembedded or only-subsetted fonts, transparency in PDF/A-1, missing XMP metadata, external hyperlinks to web resources, and unmapped Unicode text. All are fixable in the source application before re-exporting.