Logo
Back to Blog
Productivity August 13, 2026 8 min read

What Is PDF/A? The Archiving Format, Explained Plainly

O

OmniWebKit Team

Document Tooling

Share:
Article Cover Image

PDF/A is a PDF that carries everything it needs inside itself. No borrowed fonts, no external colour profiles, nothing that runs. That is the whole idea, and every rule in the standard follows from it.

The problem it solves is quiet and slow. You open a contract from 2004 and the text has reflowed, the typeface is wrong and a signature block now sits on the following page. Nothing corrupted the file. It simply asked for a font the reader could not supply, so the reader guessed. On a brochure that is irritating. On a document where a page break decides which clause sits above a signature, it is a legal problem. Making the file self-contained removes the guessing entirely.

The PDF/A vs PDF difference, in one sentence each

A normal PDF is allowed to make assumptions about the computer opening it. A PDF/A file is not.

That single change cascades. Fonts have to be embedded rather than named. Colour has to be defined by an embedded profile rather than left to the monitor. JavaScript, embedded video and external file references are banned, because anything that reaches outside the file is a dependency that can rot. Encryption is banned too, which surprises people until they think about who will remember the password in thirty years.

What does not change is how it looks. A compliant file renders identically to the original, opens in any PDF reader, and prints the same. You are not converting to a different format so much as removing the file's ability to depend on anything.

The ISO 19005 standard explained without the jargon

PDF/A is published as ISO 19005, in parts. Each part came later and relaxed the rules a little.

  • Part 1 arrived in 2005 and is the strictest. No transparency, no layers, no attachments.
  • Part 2 allows transparency, layers and JPEG 2000 compression, plus attached files provided they are themselves PDF/A.
  • Part 3 is Part 2 with one change: it permits any file type as an attachment, which is what makes hybrid electronic invoices possible.
  • Part 4 is the newest and is not yet widely requested.

Alongside the part number sits a conformance level, usually written as a lowercase letter. Level b means basic, and covers visual reproduction — the file will look right. Level a adds a tagged structure so software can understand the reading order, which is what accessibility depends on. Almost everything described casually as PDF/A is level b, and the distinction matters more than most people realise. We compared the parts properly in which conformance level you need.

One practical warning. Level a cannot be manufactured after the fact. The structure has to come from the program that authored the document, so a converter cannot invent it. If someone asks you for PDF/A-1a, the fix is in Word or InDesign, not in a conversion tool.

Why archives require PDF/A rather than ordinary files

In our experience the requirement almost never comes from a preference for tidiness. It comes from having been burned.

Records teams deal in decades. They have opened files where a chart lost its colours, where a scanned exhibit turned out to reference an external image that no longer existed, and where an embedded script threw an error box in front of the content. Each of those is a small failure that becomes serious when the document is evidence.

So the rule they land on is blunt: the file must be complete on its own. That is easier to enforce than a list of things not to do, and it is testable — you can check whether the fonts are inside the file without knowing anything about the document.

The honest caveat is that compliance is not preservation. A PDF/A file on a failing hard drive is just as lost as any other file. The format protects against format decay, not against storage neglect, and anyone selling it as a complete archiving strategy is overselling.

What a self-contained document format cannot do for you

Four limits worth knowing before you rely on it.

It does not make scans searchable. The standard never required a text layer. A scanned page converted to PDF/A is fully compliant and completely unsearchable, which catches people out constantly. Optical character recognition has to happen first, as we explain in turning scanned PDFs into PDF/A.

It does not make documents accessible. Level b says nothing about screen readers.

It does not survive editing. The moment you change the document it stops being a faithful archival copy, and the declaration inside it becomes a claim that is no longer true.

It can silently drop content. Conversion tools generally remove whatever the standard forbids and carry on rather than failing. Usually nothing visible goes missing. Occasionally a form button or an embedded media object does, and no one tells you. Compare the output against the original before filing anything that matters.

How to get a PDF/A file when you need one

Three routes, in rough order of reliability.

Export directly from the program that created the document. Word, LibreOffice and InDesign all offer it, and this route preserves the most information because the authoring program still knows the document structure. If you are heading for a university deposit, the specifics are in saving a Word document as PDF/A.

Convert an existing PDF. This is the common case, and what our PDF/A converter does — it embeds the fonts, standardises the colour and strips what the standard forbids.

Then check the result. We would not say this if rejections were rare, but they are not, and the most common cause is a missing colour profile that no tool warned about. Our PDF/A validator reads the file in your browser and tells you what it found.

So what is PDF/A, in the end?

A discipline more than a format. It is an ordinary PDF with every external dependency removed, so that opening it in 2050 requires nothing except an understanding of PDF itself.

Use it for anything with a long life and a fixed final state: contracts, filings, theses, records. Skip it for working documents, forms and anything you expect to revise. And if you have inherited a PDF/A file that you now need to change, that is a normal situation with a normal answer, covered in editing a PDF/A file.

Frequently Asked Questions

Does a PDF/A file need special software to open?

+
No. Any PDF reader opens it, because a PDF/A file is a PDF with extra restrictions rather than a different format. What you may notice is your reader switching into a protected mode that blocks edits. That is the reader recognising the archival flag, not the file being locked.

Is PDF/A a compressed or a smaller format?

+
The opposite, usually. Embedding every font the document uses adds weight, so archival copies tend to be larger than the original. A document with several typefaces, or one using a Chinese or Japanese font, can grow substantially. That size is the cost of the file being self-contained.

Can a PDF/A file contain a video or a fillable form?

+
Video, audio and 3D objects are forbidden in every part of the standard. Fillable forms are a grey area: the fields can survive, but many converters flatten them into static text. If a form matters, fill it in before archiving rather than after, because you may not get a second chance.

Does PDF/A guarantee my document is legally valid?

+
No, and the two get conflated constantly. The standard governs whether the file renders faithfully in future, not whether its contents carry legal weight. Signatures, retention periods and admissibility are separate questions governed by whoever is receiving the document.

Why does the standard forbid passwords when security matters?

+
Because the threat model is time, not intruders. A password guarantees that in thirty years, when everyone who knew it has retired, the document becomes unreadable. Archives protect files with access controls on the storage system instead, where the keys can be managed and handed over.

What happens if I open a PDF/A file in twenty years and the format is obsolete?

+
That is the scenario the design anticipates. Because the file carries its own fonts and colour definitions and contains nothing executable, a future reader needs only to understand the PDF syntax itself, which is publicly documented. It is a bet on documented simplicity rather than on any vendor staying in business.

Is PDF/A the same thing as a tagged or accessible PDF?

+
No. Accessibility lives in conformance level a, which adds a structure tree, a reading order and alternative text. Most files described as PDF/A are level b, which covers visual reproduction only. A perfectly valid level b document can be unusable with a screen reader.

Who actually decides that my document has to be PDF/A?

+
Whoever receives it. Court registries, university repositories, tax authorities and corporate records teams each set their own rule, and the specific part they accept varies. Always read the requirement rather than assuming, because a valid file at the wrong level gets rejected just as firmly as an invalid one.

If PDF/A is so much better, why is it not the default for everything?

+
Because most documents are not archives. The restrictions that make the format durable also strip out the interactivity that makes an ordinary PDF useful, and the file gets bigger in exchange. For a form, a presentation or anything with a two-week lifespan, that is a bad trade.

Can I tell whether a file is PDF/A just by looking at it?

+
Not visually, because a compliant file looks identical to the original. Some readers show a blue notification bar when they detect the archival flag. The reliable answer is to read the file metadata, which is what a validator does in a couple of seconds.

Tags

#PDF#PDF/A#Archiving#Compliance