Adding tags to an existing PDF
Old templates, OCR’d scans, files straight from prepress: plenty of PDFs were never tagged. A retroactive base tree at least gives a screen reader a way into the document.
Last updated:
What tags do in a PDF
Tags form an invisible structure layer. They state the role of each piece of content (heading, paragraph, list, table, figure, link) and together make up the structure tree. A screen reader uses that tree to read the document in a sensible order and to jump from heading to heading. Without tags it sees only a stream of characters and drawing commands. At worst, it reads out nothing or nonsense.
PDF/UA (ISO 14289) requires a fully tagged PDF. To check if your document already has tags, see Check PDF/UA online.
Fix the source first
Tags are most reliable when they come from where the structure is known. Word, InDesign and other authoring tools understand heading styles, real lists, tables and alternative text, and write them into the PDF as tags on export. If you have the source file, a clean export beats almost any repair after the fact.
Retroactive tagging is for the cases where the source is gone, the export produced no tags, or a large archive of legacy PDFs is waiting.
What EasyPDFUA adds when tagging
The tagger is deliberately conservative. It only applies changes that can be derived unambiguously from the existing PDF and writes them to a new copy. The original is never modified.
Each page receives a generic P tag with marked content, and MarkInfo flags the document as tagged. Title and language go into the XMP metadata, the DocInfo title is set and DisplayDocTitle switched on. The copy also gets the pdfuaid:part = 1 entry. That entry declares PDF/UA-1 as the target profile. It is not a passed validation.
Where a web or email address stands alone on a line of text, EasyPDFUA creates a real /Link annotation together with a Link structure element, an OBJR and a ParentTree entry. Pages with such links get /Tabs /S so the tab order follows the structure. Addresses in the middle of longer lines are left alone.
For fonts, the tagger fills in missing ToUnicode mappings when they can be derived unambiguously from the embedded TrueType data or through an exact glyph match against a locally installed reference font. A missing TrueType font program is embedded only if a local font file has the same PostScript name, its licence bits permit embedding and its widths match the PDF widths exactly. A font that merely looks similar is never substituted.
Encrypted, signed or already tagged PDFs are rejected rather than changed at risk.
What the base tree does not cover
A retroactively generated base tree does not make a PDF accessible. EasyPDFUA does not create heading, list, table or figure tags and does not write alternative text. That takes an understanding of the content. Is the bold line a heading or just emphasis? Which cells are table headers? What does the chart show?
This work is done in Adobe Acrobat Pro, axesPDF or CommonLook, or by exporting the source file again. Where the requirements of PDF/UA and WCAG differ is covered in PDF/UA vs. WCAG.
The workflow in the application
- Open the application and select a PDF (up to 5 MB for free, up to 50 MB with Premium).
- Check or enter the document title and language. Both are written into the tagged PDF.
- Choose Run tagging. If no safe change is possible, the report gives the reason.
- Download the copy. It expires after 15 minutes.
Every generated file should then go through PAC or veraPDF and a manual pass. The usual messages are explained in Common PAC and veraPDF errors. An EasyPDFUA result does not certify full PDF/UA conformance.
Frequently asked questions
Can I add tags to a PDF for free?
Yes. The free plan tags PDFs up to 5 MB without any payment details.
Is a retroactively tagged PDF automatically accessible?
No. It then has a technical base structure. Headings, lists, tables and alternative text have to be added by hand or come from an accessible export of the source file.
Will my original PDF be changed?
No. EasyPDFUA writes every change to a new copy.
Why was my PDF rejected?
Usually because it is encrypted, signed or already tagged, or because its structure is technically ambiguous. A change could then invalidate the signature or damage existing structures.
Does tagging work with scanned PDFs?
Not without preparation. A scan without text recognition contains only images. Once OCR has produced real text, tagging makes sense.