PDF/A: The ISO 19005 Standard for Long-Term PDF Archiving

PDF/A is a constrained subset of the regular PDF format, first published as ISO 19005-1 in 2005. A normal PDF is flexible in ways that are dangerous for archiving: it can depend on fonts installed on the reader's computer, link to external files or content, be encrypted, or run scripts. Any of those can cause a document to render incorrectly, or not at all, on a viewer that does not yet exist.

The core problem PDF/A addresses is long-term digital preservation. Archives, courts, and regulators must keep records reliable for decades, long after the software that created them is gone. PDF/A removes the risky features and requires every file to be fully self-contained and unambiguous, so the visual appearance can be reproduced faithfully far into the future without any external resource.

The standard has grown in parts, and later parts do not replace earlier ones. Each remains a valid ISO standard, so you choose the version that matches your needs and your recipient's requirements.

In practice, PDF/A-1 is the most rigid and widely accepted baseline, PDF/A-2 relaxes several technical limits, and PDF/A-3 is the basis for hybrid documents that carry their own machine-readable source data, such as European electronic invoices (Factur-X / ZUGFeRD).

Within a version, a conformance level states how much the file guarantees. Level b (basic) is the minimum: it ensures the document's visual appearance can be reliably reproduced. Level a (accessible) is the strictest: it adds a full tagged logical structure plus Unicode character mapping, so the text can be searched, extracted, reflowed, and read by assistive technology.

Level u (Unicode) was introduced in PDF/A-2 as a middle ground: it requires that all text be mapped to Unicode, but does not require the full tag structure of level a. PDF/A-1 offers only levels a and b, while PDF/A-2 and PDF/A-3 add level u. PDF/A-4 drops the a/b/u naming, making Unicode-mapped text mandatory for every file and adding variants for embedded files (PDF/A-4f) and engineering content (PDF/A-4e).

Every conformance level shares a set of hard rules aimed at self-containment and device independence. A file can look like a PDF/A and still fail these, which is why validation matters.

Color handling is a common stumbling block. If the document uses device-dependent color such as DeviceRGB or DeviceCMYK, it must include an output intent with an embedded ICC profile so future viewers reproduce the exact colors rather than guessing.

Many applications export PDF/A directly. Word processors and Acrobat offer a Save as PDF/A option, LibreOffice exports it from its PDF dialog, and tools like Ghostscript and preflight engines can convert an existing PDF, embedding fonts and flagging what they cannot fix automatically. Scanned paper should be run through OCR first, so the archive contains real, Unicode-mapped text rather than a flat image — this is essential for levels a and u and for future searching.

Creating a file is not the same as producing a conformant one. Validate before you archive. veraPDF is the open-source, industry-supported reference validator, developed by the Open Preservation Foundation and the PDF Association and originally funded through the EU PREFORMA project; it checks a file against every clause of the standard and produces a machine-readable report. Acrobat's Preflight also reports PDF/A errors. A file merely labeled PDF/A is not guaranteed to be valid PDF/A until a validator confirms it.

PDF/A is most often a compliance requirement rather than a choice. Any organization with legal retention obligations or a mandate to preserve records for the long term is a candidate, and many bodies specify it explicitly for submissions.

When a recipient requires PDF/A, they usually name the exact part and level, so confirm the target (for example PDF/A-1b or PDF/A-2u) before converting.