PDF/A: what it is and what it is for

1What is PDF/A and what is it for?

PDF/A is a family of rules, set out in the ISO 19005 standard, that restricts PDF so a document can be preserved and displayed the same way for decades. An ordinary PDF can depend on fonts installed on your computer, on resources downloaded when it opens, or on features that some programs eventually stop supporting. PDF/A removes those dependencies.

It is used in historical archives, libraries, court and administrative records, accounting and any setting where you must prove that a document is kept exactly as it was saved. It does not improve visual quality or make the content more secure: its value is long-term predictability.

2How does a PDF/A work under the hood?

The core idea is that the file is self-sufficient. Every font it uses must be embedded, so the text looks the same even if that typeface does not exist on the computer of whoever opens it. Colors must be defined in a way that does not depend on the screen or the program, usually through a color profile included in the file.

The file also declares in its metadata that it meets a specific part and level of the standard, and anything that could stop working is prohibited: encryption, JavaScript, embedded audio and video, and references to external content. That declaration is not enough on its own; to know whether the file truly complies, you have to run it through a validator such as veraPDF.

3What is the difference between PDF/A-1, A-2, A-3 and A-4?

The parts of the standard appeared over time, and each builds on a different version of the PDF format. Within A-1, A-2 and A-3 there are also conformance levels: B (basic, guarantees visual appearance), A (accessible, requires tagged structure) and U (in A-2 and A-3, requires that all text can be mapped to Unicode).

• PDF/A-1 (ISO 19005-1, 2005): based on PDF 1.4; the most restrictive, with no transparency or layers.
• PDF/A-2 (ISO 19005-2, 2011): based on PDF 1.7; allows JPEG 2000 compression, transparency and layers.
• PDF/A-3 (ISO 19005-3, 2012): like A-2, but allows embedding files in other formats, such as an XML file or the original spreadsheet.
• PDF/A-4 (ISO 19005-4, 2020): based on PDF 2.0, with a different level organization than the earlier parts.

4When should you use PDF/A, and when not?

Use it when a regulation, a government agency, an archive or a client expressly requires it, or when you truly need the document to stay readable many years from now: long-lived contracts, case files, technical reports, theses or accounting records. In those cases, generate the PDF/A at the source and validate it before handing it over.

It is not a good fit for temporary documents, for files that need an open password, or for pieces with interactivity and multimedia. Requesting it out of habit makes no sense either: each organization specifies which part and level it accepts, and submitting a different one can be rejected just like an ordinary PDF.

5Common mistakes and myths about PDF/A

One myth is believing a PDF/A is a higher-quality or more compressed PDF. It is not: it can even be larger, because it embeds fonts and color profiles. Another is thinking PDF/A guarantees the content is authentic or signed; it only ensures the way the file is preserved, not the truthfulness of the document.

It is also a mistake to assume that any PDF that looks fine is already PDF/A, or that renaming the file is enough. Conformance is technical and verifiable, and a document can look perfect and still fail validation because of a single unembedded font. Finally, PDF/A does not prevent copying or printing: it is not a protection mechanism.

6Common problems and how to fix them

The most common validation failures are fonts that are not embedded, colors with no defined profile, transparency used in a variant that does not allow it, external links or resources, and the presence of encryption. The validator tells you which rule is being broken, and the usual fix is to correct it in the source program and export again, choosing the PDF/A option.

If the organization does not specify the variant, ask about it before converting; A-1 and A-2 at level B are the most widespread, but there is no single correct answer. And if you start from a scan, keep in mind that many requirements also expect a text layer, so it is best to run OCR first.

7PDF/A and working in Rayo PDF GO

Rayo PDF GO does not create PDF/A or certify a file's conformance. If you are asked for it, generate the PDF/A with the source program or a dedicated tool and validate it. Keep in mind that any later modification can break conformance: Protect PDF adds encryption, which PDF/A prohibits, and operations such as Merge PDF, Compress PDF or Flatten PDF rebuild the file, so there is no guarantee it will still comply with the standard.

A good practice is to do all the assembly first with Rayo PDF GO, which runs on your device without uploading anything, and leave the PDF/A conversion and its validation as the last step. Always keep the original, because once it has been converted and modified, there is no going back.