Create and validate PDF/A: convert PDF to PDF/A with PDF24

by

PDF/A is a kind of PDF made for long-term archiving: the file should look exactly the same on any device, years from now. With PDF24 you create PDF/A for free – online in your browser, offline in the PDF24 Toolbox, or directly when printing with the PDF printer – and check with veraPDF whether a file really follows the rules. This page explains which level to choose, what PDF24 fixes automatically during the conversion and what to do when a validation fails.

Contents

What is PDF/A?

PDF/A is an international standard (ISO 19005). A PDF/A file contains everything it needs to be displayed and leaves out everything that could change or go missing:

  • All fonts are embedded, so the text looks the same on every computer, even where the font is not installed.
  • Colours are defined precisely by an embedded colour profile.
  • No password protection, no JavaScript, nothing that has to be loaded from outside to display the file (links to websites are allowed).
  • Metadata such as the title is stored in a fixed form, together with an identification like "I am PDF/A-2b".

You open a PDF/A file with any PDF program, just like any other PDF. Whether you need PDF/A, and which level, is decided by whoever receives the file; they usually name the level as well. If in doubt, ask them.

Which level: PDF/A-1b, 2b or 3b?

The number is the part of the standard, the letter is the conformance level. PDF24 creates level b ("basic"): it secures the visual appearance of the pages and is the minimum level of every part. If "PDF/A-2" is required without a letter, PDF/A-2b meets it.

  • PDF/A-2b – the usual choice and our recommendation if nothing else is required. Allows transparency, layers and form fields.
  • PDF/A-1b – the oldest version. Only use it when it is explicitly required: PDF/A-1 allows no transparency and no attachments. Pages with transparent content (e.g. images with a transparent background, shadows, semi-transparent shapes) are then saved as one single image, and the text on them can no longer be selected or searched. Form fields are fixed into the page.
  • PDF/A-3b – like PDF/A-2b, but any kind of file may be attached. E-invoices in the ZUGFeRD or Factur-X format are PDF/A-3 files with the invoice XML inside (more).
  • PDF/A-4 and PDF/A-4f – the newest version (no letter levels; 4f with attachments). PDF24 offers them when saving with an output profile: when printing, for a finished PDF in the PDF24 Assistant (in the Toolbox "Open PDF printer assistant", add the PDF with "+", then "Save as...") or from the command line. "Convert from PDF" and the online conversion do not offer them.

PDF24 does not create levels a (accessible, with structure tags) and u (all text as Unicode). You can still check files against these levels and against PDF/UA (accessibility) with Check PDF/A.

Which way suits you?

  • Convert a finished PDF, on any device: online with PDF to PDF/A → step by step
  • No internet, the file stays on your PC, many files at once: PDF24 Toolbox, "Convert from PDF" → Toolbox
  • The document is still open in Word, Excel or another program: save it as PDF/A right away when printing to the PDF printer → When printing
  • Every PDF should become PDF/A automatically: your own profile or automatic saving → Always PDF/A
  • Check whether a file really is PDF/A: Check PDF/A online or in the Toolbox → Validate
  • Scripts, servers, batch processing: Command line

The online tools work on Windows, Mac, Linux, iPhone and Android. The free PDF24 Creator with the Toolbox and the PDF printer is available for Windows.

Convert PDF to PDF/A online

  1. Open PDF to PDF/A.
  2. Click "Choose files" or drag one or more PDFs into the file box. If a file has a password to open it, PDF24 asks for it. PDF/A does not allow encryption; if the file is to stay PDF/A, do not protect it again after the conversion.
  3. Choose the level under "PDF/A level". "Automatic (recommended)" creates PDF/A-2b, or PDF/A-3b for files with attachments so that the attachments are kept. If PDF/A-1b, PDF/A-2b or PDF/A-3b is required, select exactly that level.
  4. Optional: tick "Remove document information" to remove title, author, subject and keywords. The file stays PDF/A.
  5. Click "Convert", then "Download".

Below the result PDF24 tells you when it did something you might not expect: "Created as PDF/A-3b because the file contains attachments." or "Attachments were removed because PDF/A-1b does not allow them." (PDF/A-2b only keeps attachments that are PDF/A themselves). Your files are deleted from the PDF24 server automatically after one hour.

The general converter Convert PDF to ... offers the same with the format PDF/A. For a Word, Excel or image file, convert it to PDF first (create PDF) and then to PDF/A – or, on your PC, save it as PDF/A directly when printing.

Offline in the PDF24 Toolbox

  1. Open the PDF24 Toolbox (part of the PDF24 Creator) and choose "Convert from PDF". Note: this is not the tool "Convert to PDF".
  2. Add one or more PDFs.
  3. Under "Format" choose "PDF/A-1", "PDF/A-2" or "PDF/A-3". The result is always level b, i.e. PDF/A-1b, PDF/A-2b or PDF/A-3b.
  4. Leave "Color" at "Do not change" unless you need otherwise. "Remove document information" works as online.
  5. Click "Convert" and save the result.

Behind the cog ("Additional options") you can enter "Title", "Subject", "Author" and "Keywords"; empty fields keep the values of the file. The "Profile" there is the colour profile of the PDF/A: "sRGB IEC61966" (default) for documents that are read on screen or printed in the office, "ISO coated v2" or "ISO coated v2 300%" for offset printing when a print shop asks for it. To have the content converted to CMYK as well, also set "Color" to "CMYK".

Files with attachments: choose "PDF/A-3" in the Toolbox – it does not switch to it by itself like "Automatic" online. With "PDF/A-1" a file with attachments fails validation, with "PDF/A-2" it passes only if the attachments are PDF/A themselves. Removing the attachments the chosen level does not allow comes with the next version.

If a PDF has a password to open it, the Toolbox asks for it. By default the Toolbox converts with "PDFBox". In the Settings under "Preferred Methods" you can choose "Ghostscript" for "Create PDF/A" instead; this page describes the default.

Save as PDF/A when printing

If the document is still open in Word, Excel, your browser or another program, create the PDF/A in one step:

  1. Print the document to the printer "PDF24" (how). The PDF24 Assistant opens.
  2. Click "Save as..." and choose the format PDF.
  3. Under "General" set "PDF standard" to the required version, e.g. "PDF/A-2". The choices are "PDF/A-1", "PDF/A-2", "PDF/A-3" (each level b), "PDF/A-4" and "PDF/A-4f".
  4. Save the file.

This way the fonts of the original are used, because they are installed on your PC – the most reliable way to a clean PDF/A, especially with unusual fonts.

More settings are on the page "PDF/A" on the left (they become available once a PDF/A standard is selected):

  • "Title": the document title that PDF/A files should have.
  • "Profile" (DestOutputProfile): the colour profile of the PDF/A. Keep "sRGB IEC61966" unless you need something else.
  • "Force PDF/A conformance": only if a file does not become valid otherwise. PDF24 may then also change content that cannot be converted any other way, e.g. fix form fields into the page or save parts as an image.

While a PDF/A standard is selected, the page "Security" shows a note: password and permissions are not applied, because PDF/A does not allow encryption.

Always save as PDF/A

Create your own profile: in the PDF24 Assistant click "Output profiles" at the bottom (in the Toolbox: "Manage output profiles"). Under PDF › "General" set "PDF standard" to the version you need, e.g. "PDF/A-2". At the bottom, next to "Profiles", click "New", enter a name such as "Archive" and click "Save". Changes to a profile are only kept when you click "Save" in this profile bar.

From the next printout on, the profile is in the list above "Save as PDF". Select it once; the Assistant remembers the choice, and "Save as PDF" then always creates PDF/A.

Without any question: in the PDF24 Settings, section "PDF Printer", choose "Automatically save documents after printed" under "PDF Printer Tool", and under "Auto Save" select your PDF/A profile as "Profile". Every printout is then saved as PDF/A without asking (more about automatic saving). To keep the Assistant for normal printouts, add a second PDF printer with "+", e.g. "PDF24 Archive", that saves as PDF/A automatically.

How to roll out such a profile to many PCs is described on PDF24 Creator silent install and deployment.

Validate PDF/A and read the report

Only a validation shows whether a file really meets the PDF/A rules. A file that claims to be PDF/A is not necessarily valid. PDF24 checks with veraPDF, the reference validator of the PDF Association:

  • Online: Check PDF/A
  • Offline: "Check PDF/A" in the PDF24 Toolbox (since version 11.32.0). The file does not leave your computer. The tool needs the component "OpenJDK (GPL)".

How it works:

  1. Add one or more PDFs.
  2. Leave "Check against" at "Automatic (what the file claims)": PDF24 checks against the level the file claims in its metadata, or against PDF/A-2b if it claims none. If a specific level is required, choose it, e.g. "PDF/A-1b". All levels from PDF/A-1 to PDF/A-4 and PDF/UA-1/-2 are available.
  3. Online, click "Check". In the Toolbox the check starts as soon as a file is added; changing "Check against" checks again.

How to read the report:

  • "The file claims: …" – the level the file says it meets. "The file claims no PDF/A or PDF/UA conformance" means: it is not PDF/A (yet).
  • "Meets PDF/A-2b" – everything is fine.
  • "Does not meet PDF/A-2b" with "… rules violated in … places". Under "Violated rules" each rule is explained in plain words, with its clause in the standard, how often it is violated and on which pages. Point at a rule to see veraPDF's exact wording.
  • "Convert to PDF/A" next to a file that fails opens the conversion with this file. Check the result again afterwards.

The version number is not the PDF/A level. If a program shows "PDF 1.7" or "PDF 1.4", that is the version of the PDF format. PDF/A-2 and -3 are based on PDF 1.7, PDF/A-1 on PDF 1.4. A file saved from Word is often 1.7 already – so the number does not tell you whether the conversion worked. Only the validation does.

What PDF24 fixes automatically

Most reasons why a PDF is not PDF/A are fixed by PDF24 during the conversion:

  • Fonts: fonts that are not embedded get embedded – only the characters used, so the file stays small. If a font is missing, PDF24 uses a suitable substitute; character spacing may shift very slightly.
  • Colours: PDF24 embeds a colour profile (sRGB by default) and maps all colours to it.
  • JavaScript and actions (e.g. launching programs), video, sound and 3D content are removed. Links to websites stay.
  • Forms: form fields stay fillable in PDF/A-2b and -3b; in PDF/A-1b they are fixed into the page. Button actions (e.g. "submit form") and dynamic XFA form data are removed.
  • Transparency: is kept in PDF/A-2b and -3b; in PDF/A-1b an affected page becomes one single image (see above).
  • Attachments: see PDF/A-3.
  • Digital signatures: converting rewrites the file; an existing signature would be invalid afterwards and is removed. So convert before you sign (more about signatures).

Text, images, links and bookmarks are kept (except on pages that PDF/A-1b saves as an image). PDF/A files are often a little larger than the original because the fonts have to be inside.

PDF/A-3: attachments and e-invoices (ZUGFeRD)

PDF/A-3 allows attachments of any kind, PDF/A-2 only attachments that are PDF/A themselves, PDF/A-1 none at all. If your PDF has attachments that must be kept, use PDF/A-3b (online, "Automatic" does that for you).

An e-invoice in the ZUGFeRD or Factur-X format is a PDF/A-3 file with the invoice data inside as XML. Simply attaching an XML to a PDF/A-3 is not enough: the XML needs a fixed name and matching metadata. The PDF24 e-invoice generator does all of this automatically:

Validate an e-invoice with Validate e-invoice; whether its PDF follows the PDF/A rules is shown by Check PDF/A.

Command line

The PDF24 DocTool converts to PDF/A without any window, e.g. from a batch file or the Windows Task Scheduler. Use -applyProfile and change the PDF standard of an existing profile with -profileParam pdf-standard:

"C:\Program Files\PDF24\pdf24-DocTool.exe" -applyProfile -profile default/good -profileParam pdf-standard a-2 -silent -outputFile "C:\Archive\contract.pdf" "C:\Inbox\contract.pdf"

All PDFs of a folder:

"C:\Program Files\PDF24\pdf24-DocTool.exe" -applyProfile -profile default/good -profileParam pdf-standard a-2 -silent -expandDirs -includeFileExtensions .pdf -outputDir "C:\Archive" "C:\Inbox"
  • pdf-standard is a-1, a-2, a-3 (each level b), a-4 or a-4f (the last two since version 11.31.0).
  • -silent shows no windows and reports errors through the exit code (0 = success).
  • Word, Excel and other files are converted straight to PDF/A with -convertToPDF and your own PDF/A profile: -profile user/Archive (the name you gave the profile when you created it).

The Toolbox tools "Check PDF/A" and "Convert from PDF" also run from the command line, as a job in a JSON file (manual, chapter 9.1.2); all profile options are in chapter 11.1.4.

Problems and solutions

The recipient rejects my PDF/A ("not a valid PDF/A")

  1. Ask which level is required. Some portals expect PDF/A-1b; then convert with exactly this level (online "PDF/A level" "PDF/A-1b", in the Toolbox format "PDF/A-1").
  2. Validate the file with Check PDF/A against exactly this level.

If the file was changed after the conversion, convert it again (see the next point).

The PDF/A is no longer valid after editing

Watermarks, page numbers, merging, redacting or editing rewrite the file. It often keeps its PDF/A identification but no longer follows the rules, e.g. because a semi-transparent watermark is not allowed in PDF/A-1. So always create PDF/A last: merge, edit and run text recognition (OCR) first, then convert to PDF/A and validate. Add a signature only after that, because converting later would remove it; check the signed file again with Check PDF/A.

With PDF/A-1 the whole page is an image and the text cannot be selected

This comes from PDF/A-1 itself: it allows no transparency, so pages with transparent content are saved as an image. Use PDF/A-2b unless PDF/A-1 is explicitly required. If it is required, remove transparent elements (shadows, semi-transparent shapes, images with a transparent background) in the source document and create the PDF again.

A font is not embedded or characters are missing

If possible, create the PDF/A directly from the original document when printing, where the fonts are installed on your PC. For a finished PDF that does not contain a font, PDF24 can only embed a substitute. With some kinds of fonts even that is not possible; then the only fix is to create the PDF again from the original.

"The file is encrypted" – PDF/A with a password

PDF/A does not allow encryption, neither a password to open the file nor permissions. So a protected PDF is not a valid PDF/A, and the two do not go together. Create the PDF/A from the source file before it was protected, e.g. when printing from the original document. When you save as PDF/A in PDF24, no password is set; if you protect a PDF/A file afterwards, it is no longer a valid PDF/A. If PDF/A is required, send the file without a password.

"Check PDF/A" says "For this function no suitable component is installed"

"Check PDF/A" needs the component "OpenJDK (GPL)", and so does converting with "PDFBox" (without Java and Ghostscript, "Convert from PDF" reports "File could not be converted"). If it was deselected during installation, run the setup of the PDF24 Creator again and select it. Or convert and validate online.

The PDF/A is much larger than the original

PDF/A must contain all fonts, which makes the file larger, especially for short documents. Reduce large images with Compress PDF. Leave "Strip output intents" unticked, otherwise the file is no longer a valid PDF/A, and validate the result.

If you are missing a feature or something does not work, write to us at team@pdf24.org or ask in the PDF24 forum.