PDF/A is a special flavor of PDF designed for long term archiving, a file format built so a document opened fifty years from now looks exactly the way it does today. While ordinary PDFs depend on fonts and features that may vanish over time, PDF/A locks everything the document needs inside the file itself. Explore the full collection of document utilities in the PDF tools hub to manage your PDFs.
Why ordinary PDFs can break over time
A regular PDF can reference fonts installed on the reader’s computer, link to external images, embed audio and video, or run JavaScript. All of that works fine today, but in twenty years, the referenced font may no longer exist, the linked image may be gone, and the JavaScript engine may behave differently. For a contract, a legal record, a thesis, or a government archive, “probably fine” isn’t good enough. PDF/A solves this by banning everything that depends on the outside world.
The PDF/A versions: A-1, A-2, A-3, and A-4
PDF/A has evolved through several versions, each based on a newer generation of the PDF specification. Knowing which one you have, or need, matters because they support different features.
- PDF/A-1 the original archival standard, based on the older PDF 1.4 specification. It forbids transparency, layers, and embedded files entirely. Very widely supported, but restrictive by modern standards. Comes in two conformance levels: “a” (tagged, accessible structure) and “b” (visual appearance only).
- PDF/A-2 based on PDF 1.7, it relaxes several A-1 limits: transparency and layers are allowed, JPEG 2000 compression is permitted, and OpenType fonts can be embedded. Also adds a “u” conformance level, which requires Unicode text mapping so text can be searched and copied reliably.
- PDF/A-3 identical to A-2 in every visible respect, except that it allows embedding arbitrary files (spreadsheets, CAD data, XML invoices) inside the archival document. This makes it popular for e invoicing workflows where the human readable PDF carries its machine readable data along.
- PDF/A-4 the newest generation, aligned with the PDF 2.0 specification. It simplifies the conformance levels and drops some legacy baggage. Support is still catching up in some tools, so A-2 remains the safest choice for maximum compatibility today.
When does archival format actually matter?
- Legal and compliance records. Courts, regulators, and auditors may require documents in an archival format so evidence can’t silently degrade.
- Academic theses and dissertations. Many universities require graduating students to submit their thesis as PDF/A so the library’s copy survives decades of format churn.
- Government and institutional archives. National archives and large organizations standardize on PDF/A for records they must keep for the long haul.
- Contracts and vital records. Anything you might need to produce, unchanged, years from now is a candidate.
For everyday documents, a flyer, a resume, this month’s invoice. PDF/A is overkill. Standard PDF is perfectly fine when the document’s lifespan is measured in years rather than decades.
How to check if a PDF is PDF/A
Most desktop PDF readers show the document’s conformance in its properties. In Adobe Acrobat, open File → Properties and look at the Description tab: a PDF/A file will list something like “PDF/A-2b” as its standard. Some readers display a blue banner at the top of PDF/A documents noting that editing is restricted, since archival files are meant to be read, not changed.
Free online validators can also check a file against the PDF/A rules and report exactly which requirements it passes or fails, useful before submitting a thesis or an archive package.
How to create a PDF/A file
Most office software can export directly to PDF/A. In Microsoft Word, use File → Save As → PDF and tick the option for ISO compliant PDF/A. LibreOffice offers PDF/A export in its PDF options dialog. Adobe Acrobat can convert an existing PDF via its Preflight tool, which also validates the result. After converting, always validate the output, a file that claims to be PDF/A but fails validation may be rejected by whoever requires it.
Worked example: archiving a university thesis
Consider a concrete archival job from start to finish. A graduate student has written a 120-page thesis in Word, using three fonts, fourteen figures, and a table of contents with internal links. Her university requires the library copy as PDF/A-2b, validated, by the end of the month. Here is what actually happens to her file, step by step.
She exports from Word twice. The ordinary PDF export produces a file of 1.8 MB that looks perfect on her machine, because her machine has the fonts. The PDF/A export, with the ISO compliant option ticked, produces a file of 2.1 MB. That is 0.3 MB larger, an increase of about 16.7% (0.3 ÷ 1.8 = 0.167), and the difference is almost entirely the fonts: all three are now embedded inside the file, along with the colour information needed to render them identically elsewhere. Nothing about the extra size is waste. It is the price of the document carrying everything it needs with it, which is the entire idea of the format.
Next she validates. The first validation report flags two problems, and both are typical. One figure was pasted as a screenshot whose colours are stored in a device dependent way, so the converter re encodes it. And the document properties are missing the structured metadata the standard requires, so she fills in the title and author fields and exports again. The second validation passes. Notice what did not happen at any point: the wording never changed, no page moved, and the table of contents links still work, PDF/A-2 allows internal links, it simply forbids dependencies on things outside the file. When she later needs a smaller copy to email to a supervisor, she keeps the validated PDF/A untouched as the archive master and makes the lightweight copy from it with the compress PDF tool, understanding that the emailed copy is for reading, not for the record. If the thesis later needs a signed approval page added to the working copy, the sign PDF tool handles that everyday copy, while the archived master stays exactly as validated.
The versions compared
The four generations differ in a handful of practical ways. This table summarises what each one allows, so you can match a requirement to a version without reading the standards documents.
| Version | Based on | Transparency and layers | Can embed other files? | Sensible choice when |
|---|---|---|---|---|
| PDF/A-1 | PDF 1.4 | Not allowed | No | A recipient explicitly requires the oldest, simplest standard |
| PDF/A-2 | PDF 1.7 | Allowed | No | Archiving almost anything modern, the safe default today |
| PDF/A-3 | PDF 1.7 | Allowed | Yes, spreadsheets, XML, source data | The archive must carry its source data inside the same file |
| PDF/A-4 | PDF 2.0 | Allowed | Yes, under its own rules | Your whole toolchain confirms support for the newest standard |
Two patterns stand out from the table. First, each generation mostly removes restrictions rather than adding features readers can see, the visible document stays a document. Second, the jump from A-2 to A-3 changes nothing you can look at; it only changes what may ride along inside the file. That is why confusing the two is so common, and so consequential when an invoice archive genuinely needs its XML data embedded or an examiner explicitly forbids attachments.
Common mistakes to avoid
- Assuming Save as PDF means PDF/A. A default PDF export is an ordinary PDF, however tidy it looks. The archival option has to be chosen deliberately in the export settings, and the result still needs validating.
- Editing the archived copy. Every edit risks breaking conformance, and a silently non conforming archive is worse than no archive because everyone trusts it. Edit the source document, export a fresh PDF/A, and replace the archive wholesale.
- Password protecting an archive. Encryption is forbidden in PDF/A, because a file nobody can open in forty years is not an archive. Protect the folder or the storage, not the archival file itself.
- Choosing A-3 when the requirement said A-2. The versions are not interchangeable upgrades. Submitting the wrong one can mean rejection even though your file is, in every visible respect, identical.
- Skipping validation because the export “worked”. Exporters produce files that claim conformance and quietly miss a requirement, an unembedded font is the classic. The validator’s report, not the export dialog, is the evidence your recipient will ask for.
A final habit ties the whole workflow together: keep three things for every archived document, the source file you can still edit, the validated PDF/A itself, and the validation report that proves it passed on the day it was archived. Storage is cheap, and in ten years that little bundle answers every question anyone can ask about the record. The full set of utilities for the non archival copies lives in the PDF tools hub, so everyday sharing never has to touch the master.
When a conversion refuses to cooperate
Most PDF/A conversions succeed on the first or second attempt, but a stubborn minority fail for reasons worth recognising in advance. The most common is a font that cannot legally be embedded: some fonts are licensed for use on screen and in print but not for redistribution inside a file, and a compliant converter must refuse rather than embed them anyway. The cure is to change the document to a font that allows embedding and export again, the text is untouched, only its clothing changes. Scanned documents cause a different problem: a page that is a photograph of text contains no real text at all, so there is nothing for the archive to preserve in searchable form until optical character recognition has been run over it, and the OCR quality then becomes part of the record’s quality.
The practical response to a failed conversion is to treat the validator’s report as a worklist rather than a verdict. Each flagged item names a specific requirement and, usually, the page or object that violates it. Work through the list, fix the source document where the problem originates, and re export, editing the exported PDF directly almost always trades one violation for another. And keep the report from the failed attempt alongside the final one: together they show a recipient exactly what was wrong and what was done about it, which is the kind of paper trail archives exist to create.
Frequently asked questions
Is PDF/A better than regular PDF?
Not better in general, better for a specific job. PDF/A guarantees long term visual stability by embedding all fonts and banning external dependencies, JavaScript, and encryption. But those same restrictions make it worse for interactive documents: no forms with scripts, no embedded video, no password protection. Use PDF/A for archiving, regular PDF for everything else.
Can I convert a regular PDF to PDF/A?
Yes, using tools like Adobe Acrobat’s Preflight or office suite export options. Conversion embeds the fonts and strips non compliant features. Some documents can’t convert cleanly, for example, a PDF that uses a font you don’t have licensed for embedding, in which case the converter will flag the problem for you to fix.
Which PDF/A version should I use?
Unless someone specifies otherwise, PDF/A-2b is the safe default: it supports modern features like transparency and is accepted virtually everywhere. Use PDF/A-1b only if a recipient explicitly requires the older standard. PDF/A-3 makes sense when you need to embed source files like spreadsheets. PDF/A-4 is the future but has the least tool support today.
Can a PDF/A file be edited?
Technically yes, but editing breaks the point: most editors warn you that saving changes may invalidate the PDF/A conformance. The archival workflow is to keep the PDF/A as the untouched record and edit the source document, then export a fresh PDF/A.
Does PDF/A support digital signatures?
Yes, digital signatures are permitted in PDF/A and are commonly used on archived legal documents. The signature must follow the archival rules (no reliance on features added after signing), and validators check this. A signed PDF/A remains verifiable decades later, which is precisely the point of the format.