What Is PDF/A? (And When You Actually Need It)
PDF/A documents are the archival cousin of regular PDF — same file, built to stay readable for decades. Plain-English guide and how to convert in 5 seconds.

You sent a contract as a PDF, the recipient replied:
"Please send it as PDF/A-2b for our records."
Don't panic. PDF/A documents are just regular PDFs with the unstable parts removed. Here's the short explanation, when you actually need them, and how to convert.
What "regular" PDF looks like under the hood
A normal PDF can include:
- Text and images (the obvious stuff)
- Embedded fonts — but the file can also point to fonts on the viewer's computer. If those fonts go missing, the text shows in a substitute font.
- External media — video, audio, 3D models.
- JavaScript — yes, PDFs can run code. Forms with calculations, etc.
- Encryption — password-locked PDFs.
- Transparency, layers, hidden content — fine on modern viewers, but renders differently on old ones.
Most of that is fine for sending a CV in 2026. None of it is fine for sending a document a court will need to read in 2050.
What PDF/A removes (and why)
PDF/A enforces a few simple rules:
- All fonts must be embedded. The file carries its own fonts. Open it on a stone tablet in 2080 — fonts still work.
- All colour spaces must be explicit. No "use the system's default colour profile". Colour means the same in 50 years.
- No external references. No links to fonts on a server, no embedded video that needs Flash, no anything that requires the outside world.
- No JavaScript. No active code. Just a document.
- No encryption. Archives need to stay readable, not locked away behind a password somebody will forget.
- No transparency or layers (in older PDF/A-1; relaxed in PDF/A-2 and PDF/A-3 — see below).
The result: a PDF that opens identically in 2026, 2040, 2070. Same fonts, same colours, same layout.
The variants you'll hear about
The format is an ISO standard (ISO 19005), and the US Library of Congress keeps plain-English format descriptions of PDF/A-1 and PDF/A-2 if you want the primary source behind this table.
| Variant | Year | What's new |
|---|---|---|
| PDF/A-1a | 2005 | Strictest. Requires text to be **tagged** (accessible to screen readers). For long-term + accessibility. |
| PDF/A-1b | 2005 | Strict, no tagging required. Most common everyday choice. |
| PDF/A-2 | 2011 | Adds JPEG2000, transparency, layers. Smaller files, more modern visuals. |
| PDF/A-2b | 2011 | Most common for **business and legal** today. |
| PDF/A-3 | 2012 | Allows embedded attachments (Excel, XML inside the PDF). Used in e-invoicing. |
| PDF/A-4 | 2020 | Cleanest, modern revision. Slowly being adopted. |
If you're told just "PDF/A" with no number, PDF/A-2b is the safe default. It's modern enough to be sensible, common enough to be widely supported.
When you actually need PDF/A
Specific cases:
- Court filings (US PACER, EU courts, UK Crown Court) often require PDF/A.
- Tax / accounting records must be retained 7-10 years; PDF/A is the standard.
- Academic theses — many universities require PDF/A submission.
- Government archives (US National Archives, UK National Archives, German Bundesarchiv) standardize on PDF/A.
- Medical records in many jurisdictions.
- E-invoicing — the ZUGFeRD / Factur-X standards in Europe use PDF/A-3 with embedded XML.
If none of those apply to you, you probably don't need PDF/A. Regular PDF is fine for everyday sharing.
How to convert a PDF to PDF/A — step by step
Open the PDF/A tool
Go to pdfty.com/tools/pdf-to-pdfa.
Upload your PDF
Drag in the existing PDF. Free up to 20 MB.
Pick a variant
Default is PDF/A-2b — the safe choice whenever nobody specified a variant. If your recipient specified a different variant (1a, 1b, 3b), pick that.
Convert
We embed fonts, fix colour spaces, strip JavaScript and external references. Our 20-page, 21 KB test file took 0.1 s; large or image-heavy files take longer.
Download
The result is a PDF/A-compliant file. Viewers (Adobe Reader, Preview, browsers) all open it normally — the difference is invisible to readers but huge for archivists.
What the conversion actually did to our test file
So you can see the difference rather than take our word for it, we ran the PDF/A-2b conversion our worker runs (Ghostscript 10.02.1, inspected with pikepdf 9.0.0, 2026-10-06) on a 20-page text-only report that used the standard Helvetica font without embedding it — exactly the "points to a font on the viewer's computer" case described above.
- Fonts: the input's
/Helveticawas not embedded. The output carries an embedded subset,/CSYVCN+Helvetica— the file now ships its own glyphs. - Identification: the output's XMP metadata declares
pdfaid:part = 2andpdfaid:conformance = B, which is how a validator recognises a PDF/A-2b claim. - Size: 21 KB in, 22.6 KB out — about 8% larger, all of it font data.
- Time: 0.1 s.
Your file will behave differently: a PDF whose fonts were already embedded will barely grow, and a scan with large images will take longer to process. Treat these as one data point, not a rule.
Frequently asked questions
Will my PDF look any different after conversion?
No — visually it's identical. PDF/A is about what's inside the file (self-contained fonts, explicit colour profiles, no external refs), not what the page looks like.
Will the file size change?
Usually it gets slightly bigger, because we embed all fonts into the file. In our own test, a 21 KB report with one unembedded font grew to 22.6 KB — about 8%. For PDFs that already had everything embedded, the size barely changes; a document with many unembedded fonts grows more.
Can I still password-protect a PDF/A?
No — PDF/A explicitly bans encryption. The whole point is that archives should stay openable for decades, and a password is a single point of failure. If you need both archival and confidentiality, store the PDF/A inside an encrypted folder (your OS handles that) instead.
Will my hyperlinks survive?
Internal hyperlinks (between pages of the same document) — yes. External hyperlinks (to websites) — they're preserved as text, but PDF/A-1 bans clicking them. PDF/A-2 allows them. So pick PDF/A-2b if clickable links matter.
What if my document has video, audio or JavaScript?
It gets stripped during conversion. That's the point of PDF/A — only the document itself, no active content.
Will it work with scanned PDFs?
Yes — scanned PDFs convert to PDF/A perfectly (an image inside a PDF is already self-contained). If you also need the scan to be searchable later, run OCR before conversion.
How do I open PDF/A documents?
Exactly like any other PDF — no special software needed. PDF/A documents open in Adobe Acrobat Reader, Apple Preview, Chrome, Edge, and every phone PDF app, because a PDF/A file is still a valid PDF underneath. The only visible difference: Acrobat shows a blue banner ("This document claims to be PDF/A...") and opens the file in read-only mode by default, to discourage accidental edits to an archival copy. You can switch that off in Acrobat's preferences if it annoys you. Your recipients don't need to do anything special either — that's the point of the format.
How do I verify a file is really PDF/A?
Open it in Adobe Acrobat Reader (free). The blue banner at the top will
say "This document claims to be PDF/A-2b." Note the word claims: the
banner only reads the pdfaid metadata tag. For real verification, run
the file through veraPDF, the open-source
validator developed with the PDF Association, which checks every rule of
the standard and reports each failure.
What's the difference between PDF/A and PDF/X?
PDF/X is for printing (commercial print houses, magazines). It enforces specific colour profiles and resolution. PDF/A is for archives. They're cousins, not interchangeable.
The pdfty team builds privacy-first online PDF tools — compress, convert, OCR, sign and protect. Files are deleted within 1 hour. About us →


