PDFly
Back to blog

PDF/A: why and how to archive your documents for the long term

PDF/A guarantees your documents stay readable for 10, 20, 50 years. Differences with regular PDF, PDF/A-1 to PDF/A-3 levels, and practical conversion.

May 06, 20264 min readBy PDFly

You scan a lease agreement, a notarial deed, or payslips to keep them. A few years later, you open the file — and find fonts that no longer display correctly, missing images, a broken interactive layer.

That's exactly the problem PDF/A solves since 2005. Here's what to know before archiving important documents.

PDF/A in one sentence

PDF/A (ISO 19005) is a PDF variant designed to guarantee perfect readability of a document over the very long term, regardless of the software or system that opens it 30 years from now.

Concretely, a PDF/A is a PDF with additional constraints:

  • All fonts must be embedded (no reference to a system font)
  • No external content (no URL-linked images, no JavaScript, no active forms)
  • No encryption (the file must be openable by any reader)
  • Standardized metadata (XMP, RDF) for archive classification
  • Closed color spaces (no proprietary ICC profile)

In short: a PDF/A is self-contained. Everything it needs to display correctly is in the file.

Why a "regular" PDF isn't enough

A regular PDF can reference:

  • A system-installed font — which may not exist in 20 years
  • An external link to a third-party PDF — whose URL will have changed
  • An embedded Flash plugin — whose format died in 2020
  • A specific color profile from a 2003 printer

When you open it in 2050, the application will display "default font" instead of the original, the image will be missing, the form will no longer work.

For a temporary marketing document, no problem. For a decade-retention contract, notarial deed, or accounting archive, that's unacceptable.

PDF/A levels: 1, 2, 3

Three main families, each with two conformities (a and b).

PDF/A-1 (2005)

The original standard. Doesn't support transparency, layers, or modern OpenType fonts.

  • PDF/A-1a: strict, accessible (Tagged PDF — useful for screen readers)
  • PDF/A-1b: visual conformity only (rendering will be identical, but not necessarily accessible)

PDF/A-2 (2011)

Supports JPEG2000, transparency, layers, and embedding PDF/A inside PDF/A.

  • PDF/A-2a: strict + accessible
  • PDF/A-2b: visual only
  • PDF/A-2u: visual + guaranteed-searchable text (Unicode)

PDF/A-3 (2012)

Like PDF/A-2 but allows embedding arbitrary files (CSV, XML, original Word). Heavily used for electronic invoices: visual PDF + structured XML in a single file.

  • PDF/A-3a / 3b / 3u (same nuances as PDF/A-2)

Which to choose?

Use caseRecommended level
Simple archive (contract, quote)PDF/A-1b
Archive with accessibility (public admin)PDF/A-1a or PDF/A-2a
Modern PDF with transparency/layersPDF/A-2b
Electronic invoice (Factur-X, ZUGFeRD)PDF/A-3
Academic publicationPDF/A-2u or PDF/A-3u

For 95% of cases, PDF/A-2b is the ideal compromise: modern, broad compatibility, reasonable size.

When PDF/A is legally mandatory

A few European contexts where PDF/A is required:

  • French public archives: required by SIAF (Interministerial Service of French Archives)
  • Dematerialized public tenders: most platforms accept only PDF/A
  • Factur-X / ZUGFeRD (FR + DE electronic invoice standard): mandatorily PDF/A-3 based
  • Commercial court registries: annual accounts deposits in PDF/A
  • Probative archiving (eIDAS, AFNOR NF Z42-013 certification): PDF/A-1 or PDF/A-2 mandatory for legal value

How to verify a PDF is really PDF/A

Three methods:

1. XMP metadata

Open the PDF with a metadata-capable editor (Adobe Acrobat, or Sumatra). PDF/A files declare in their metadata:

pdfaid:part = "2"
pdfaid:conformance = "B"

2. veraPDF (free open-source validation)

veraPDF is the reference tool to validate PDF/A conformity. Open-source, multiplatform, required by professional archivists.

3. PDFly (simple verification)

When converting to PDF/A on PDFly, the generated file is validated by the engine — you have the guarantee that the output is conformant.

How to convert a regular PDF to PDF/A

Several options depending on context:

Online — PDFly Premium

  1. Go to pdfly.eu/en/tools/pdf-to-pdfa
  2. Drop your PDF
  3. Choose the level (PDF/A-1b, 2b, 2u recommended)
  4. Download the result

This is a Premium feature (€4.99/month) because conversion requires server processing (embedding missing fonts, ISO verification). The file is deleted immediately after download.

Local software — Adobe Acrobat

File → Save As → choose "PDF/A" in dropdown. Included in Adobe Acrobat Pro (≈ €19/month).

Command line — Ghostscript

Free and powerful for advanced uses.

gs -dPDFA=2 -dBATCH -dNOPAUSE -sDEVICE=pdfwrite \
   -sOutputFile=archive.pdf source.pdf

Advantage: scriptable, ideal to convert 1,000 files in batch.

Common pitfalls to avoid

  1. Converting an already-signed PDF to PDF/A — the digital signature can be invalidated. Convert first, sign after.
  2. PDF/A with interactive content — JavaScript forms break on conversion. Save the active form separately.
  3. PDF/A without validation — always validate with veraPDF before legal archiving.
  4. PDF/A-3 for simple archive — overkill, choosing A-1b or A-2b is enough in 95% of cases.

In summary

PDF/A isn't an administrative gadget: it's a technical guarantee of readability over 30 or 50 years, legally required in several European contexts.

For most professional archives: PDF/A-2b is the ideal modern standard. Conversion takes less than a minute, but the benefit runs for decades.

Convert your PDFs to PDF/A now with PDFly Premium — no commitment, EU-hosted.

Mentioned tools