PDF/A Explained: The Format Built for Long-Term Archiving

PDF/A is an ISO-standardized version of PDF built for long-term archiving. It strips out anything that could stop a file from opening or looking the same decades from now: external fonts, linked media, JavaScript, and encryption. If you need a document to be readable in 2050 exactly as it looks today, PDF/A is the format the standard was written for.

What PDF/A is and the problem it solves

A normal PDF can quietly depend on things that live outside the file. It might reference a font installed on your machine, pull an image from a URL, or run a script. Open that same file in ten years on a different computer and the font falls back to something ugly, the linked image is a broken box, and the script does nothing. The document you saved is not the document you get back.

PDF/A (ISO 19005) fixes that by requiring the file to be self-contained and deterministic. Everything needed to render the page lives inside the file. There is no “it depends on your system” clause. That is the whole point of pdf long term preservation: the file carries its own future.

This matters most for records that legally have to survive. A signed contract, a court filing, a tax return, a medical record, a land title. These outlive the software that made them. PDF archiving as a discipline exists because organizations learned the hard way that a plain PDF is not a guarantee.

PDF/A vs regular PDF: the key restrictions

The difference between pdf/a vs pdf is a list of things PDF/A does not allow. Regular PDF is permissive. PDF/A is strict on purpose, and the restrictions are what buy you reliability.

Here is what PDF/A forbids or requires compared to a standard PDF:

• No external content. Fonts must be embedded. Images must be inside the file. No streaming, no remote references.
• No JavaScript or executable code. A document should display, not run.
• No encryption. An archived file must be openable without a password or a key that could be lost.
• No audio, video, or 3D objects (in the older sub-types).
• Required metadata. Document properties must be stored as XMP so a catalog system can index the file.
• Device-independent color. Color has to be defined so it renders the same on any display or printer, not “whatever your monitor guesses.”

None of these make the document look different today. They remove every way it could look different tomorrow.

Self-contained files (no external fonts or links)

Font embedding is the restriction people hit first. If you use a font the reader does not have, a normal PDF substitutes a lookalike and your careful layout shifts. PDF/A embeds the actual font glyphs into the file, so the text renders identically everywhere. That single rule is responsible for a lot of the size difference we cover below.

The same logic applies to links to external resources. A hyperlink to a website is allowed as navigation, but the file cannot depend on that URL to display its own content. Everything the page needs to draw itself is baked in.

PDF/A-1, A-2, A-3 and conformance levels (a, b, u)

PDF/A comes in sub-types, and each one relaxes or adds something. Picking the right one is most of what pdf/a compliance means in practice.

• PDF/A-1 (2005, based on PDF 1.4) is the strictest and most widely accepted. If an archive tells you “PDF/A” with no version, they usually mean this. It does not allow transparency or layers.
• PDF/A-2 (2011, based on PDF 1.7) adds JPEG 2000 compression, transparency, layers, and lets you embed other PDF/A files. Better compression, still fully archival.
• PDF/A-3 is identical to A-2 but allows one extra thing: embedding any file type as an attachment (for example, the original spreadsheet or an XML invoice next to the human-readable page). This is what electronic invoicing standards like ZUGFeRD and Factur-X use.

On top of the sub-type, there is a conformance level:

• Level b (basic): the file looks right. Visual reproduction is guaranteed, nothing more.
• Level a (accessible): everything in level b, plus a tagged structure tree and Unicode text mapping, so the document is searchable and works with screen readers.
• Level u (Unicode): a middle ground added in A-2. All text maps to Unicode (so it is searchable and copyable) but without the full accessibility tagging of level a.

So “PDF/A-2u” means PDF/A version 2, Unicode conformance. Most everyday archiving lands on A-1b or A-2b. Reach for level a when accessibility is a legal requirement.

Who requires PDF/A (courts, archives, government)

Plenty of institutions mandate PDF/A for anything they keep, and they will reject a submission that fails validation.

Courts are a common one. The U.S. federal court e-filing system and many national court systems require PDF/A for filed documents. National archives and libraries (the U.S. National Archives, the German Bundesarchiv, and many others) standardize on it for digital preservation. Government agencies specify it for records that must be retained for years or permanently.

In business, the electronic invoicing rules spreading across the EU lean on PDF/A-3 because it can carry a machine-readable XML invoice inside a human-readable page. If you deal with tax or procurement in Europe, you have probably touched a PDF/A-3 file without knowing it. When a specification says “archival PDF” or cites ISO 19005, that is your cue.

Why PDF/A files are often larger, and what you can still compress

PDF/A files tend to be bigger than the same document saved as a regular PDF, and the reason is right there in the rules. Embedding every font adds weight. A full font can add tens or even hundreds of kilobytes, and if the file uses several fonts, that stacks up. The required color profiles and XMP metadata add a little more.

That does not mean you are stuck with a bloated file. The archival rules constrain fonts and metadata, not your images, and images are usually where the real weight is. A scanned page saved at 600 DPI is enormous; the same page at 200 to 300 DPI is a fraction of the size and still perfectly legible. If your PDF/A is huge, the pictures are almost always the cause, which is exactly what we dig into in why a PDF gets so big.

You can shrink the image payload before or after conversion without breaking compliance, as long as fonts stay embedded and the color profile is preserved. Our image compressor is a quick way to bring down oversized images before you build the document, so the archival copy starts lean instead of ballooning. Just keep the resolution high enough that text in a scan stays sharp.

How to check if a PDF is PDF/A compliant

You cannot tell a PDF/A from a regular PDF by looking at the page. You have to check the metadata. There are three reliable ways.

First, open it in a PDF reader. Adobe Acrobat and many others show a blue notification bar reading “You are viewing this file in PDF/A mode” when the file declares itself compliant. That flag lives in the XMP metadata as a pdfaid entry.

Second, validate it properly. A declared flag is not proof the file actually conforms. The open-source veraPDF validator is the industry reference for real pdf/a compliance checking; it parses the file against the ISO spec and reports exactly which rules pass or fail. Archives and courts run this kind of validation on what you submit.

Third, inspect the metadata yourself. The pdfaid:part and pdfaid:conformance fields tell you the sub-type and level (for example, part 2, conformance “b”). If you want the full picture of what else is stored in there, see what PDF metadata actually contains.

Common conversion pitfalls

Converting to PDF/A is usually one export step, but a few things trip people up.

Missing fonts. If a font cannot be legally embedded (some commercial fonts block it) the conversion fails or substitutes a different font. Use fonts that permit embedding, or the file will not validate.

Transparency in PDF/A-1. A-1 does not allow transparency or layers. If your design uses drop shadows or blend modes, either flatten them first or target A-2 instead.

Losing searchable text on scans. Converting a scanned image to PDF/A-1b gives you a picture of text, not text. If you need it searchable, run OCR and target a level u or a conformance so the characters map to Unicode.

Assuming Word or a spreadsheet is a good archive format. They are not stable long-term. If you are weighing document formats for keeping, not just editing, PDF versus Word for documents you keep walks through why PDF wins for archiving.

Convert, then validate with veraPDF. Do not trust the export dialog’s promise until the validator agrees.

FAQ

What is PDF/A in simple terms? It is a version of PDF that has to stand on its own. Everything the file needs to display correctly (fonts, color, images) is stored inside it, and anything that could break over time (external links, scripts, encryption) is banned. That makes it safe for long-term archiving.

Is PDF/A better than PDF? Not better, different. PDF/A trades flexibility for reliability. For a document you will edit, share, and discard, a regular PDF is fine. For a record you must keep readable for years, PDF/A is the right choice, and often a legal requirement.

Can I convert PDF/A back to a normal PDF? Yes. A PDF/A file opens in any PDF reader like a normal PDF, and you can re-save it as a standard PDF to add features the archival format blocks, such as encryption or form scripting. You would just lose the compliance flag.

Why is my PDF/A file so large? Two reasons: every font is embedded, and scanned pages are often saved at very high resolution. The fonts you cannot remove, but you can compress the images to a sensible DPI without breaking compliance, which is usually where most of the size lives.