How to tag a PDF for accessibility
Tagging is what turns a PDF from a picture of a document into something a screen reader can navigate. This explains what tags are, how to add them in Word and Acrobat, and which parts genuinely need a human.
Not sure whether your PDF is already tagged? Check it in the browser — the result tells you immediately, along with what else is missing. Nothing is uploaded.
What a tag actually is
When you look at a PDF, you infer structure from visual cues: bigger bold text is a heading, indented text with bullets is a list, a grid of numbers is a table. None of that is recorded in the file. Visually, a heading is just text that happens to be 18-point and bold.
Tags add that missing layer. A tagged PDF contains a structure tree — a hidden outline that says “this is a level-1 heading”, “this is a list with four items”, “this is a table with a header row”. Assistive technology reads the structure tree, not the visual layout.
Without it, a screen reader gets an undifferentiated stream of text in whatever order the content happens to appear in the file, which is frequently not the order a person would read it in. Multi-column layouts are a common disaster: the text can be read straight across both columns, interleaving two unrelated sentences.
The best approach: fix the source document
This is the part most guides bury, and it is the most useful thing on this page. If you still have the original Word, Google Docs, or InDesign file, adding structure there and re-exporting is dramatically faster than tagging the PDF afterwards — often minutes instead of hours — and the structure survives every future revision.
In Microsoft Word
- Use the built-in heading styles (Heading 1, Heading 2, and so on) rather than manually making text big and bold. This is the single highest-impact step.
- Use the real list buttons for bullets and numbers, not hyphens typed by hand.
- Insert tables with the table tool, and mark the header row: select it, then in Table Properties tick Repeat as header row at the top of each page.
- Add alternative text to every image that carries meaning (right-click → View Alt Text).
- Set the document language under Review → Language.
- Fill in the title in File → Info, not just the filename.
- Run Review → Check Accessibility and fix what it reports.
- Export with File → Save As → PDF and make sure Document structure tags for accessibility is ticked under Options. Do not use Print → Save as PDF: that discards every tag.
That last point catches a surprising number of people. Printing to PDF produces a visually identical file with no structure whatsoever.
In Google Docs
Use heading styles and alt text the same way, then File → Download → PDF. Google Docs exports basic tags but handles tables and complex layouts less well than Word, so check the result.
Tagging an existing PDF in Acrobat Pro
When the source is gone, you tag the PDF directly. Acrobat Pro is the most common tool.
- Open the Accessibility tools and run Autotag Document. This produces a first draft of the structure tree. Treat it as a draft — it is frequently wrong about headings, and usually wrong about tables.
- Open the Tags panel and review what it produced. Fix mislabelled elements: body text marked as a heading, headings marked as paragraphs, decorative images that should be marked as artifacts.
- Open the Reading Order tool and confirm the sequence makes sense. This is the step people skip and the one that matters most.
- Use the Table Editor to mark header cells in every table. Autotag rarely gets this right.
- Add alternative text to each figure.
- Set the title and language in File → Properties, and under the Initial View tab set the window title to show Document Title rather than the filename.
- Run the built-in Accessibility Checker and resolve what it flags.
If the PDF is a scan
Tagging cannot begin until there is text to tag. Run OCR first (Scan & OCR → Recognise Text), then proofread the output. OCR reliably mangles proper nouns, numbers, and unusual fonts, and a screen reader will read those errors aloud exactly as recognised.
What automation can and cannot do
| Task | Automatable? |
|---|---|
| Setting document language and title | Yes, reliably |
| Generating a structure tree | Yes, as a draft |
| Running OCR on scans | Yes, with proofreading |
| Drafting alternative text | Yes, but must be reviewed |
| Deciding correct reading order | Not reliably |
| Identifying table header cells | Not reliably |
| Judging whether alt text is useful | No |
A tagged-but-wrong PDF can be worse than an untagged one, because it looks compliant to an automated check and nobody looks again. Scrambled reading order or alt text that says “chart” is a barrier dressed up as a fix.
How to verify your work
- Run the file through this checker to confirm the structural basics.
- Use Acrobat's built-in Accessibility Checker for a second automated opinion.
- Read the document with an actual screen reader — NVDA on Windows and VoiceOver on macOS are both free. Nothing else reveals a broken reading order as quickly.
- Try navigating by headings alone. If you cannot, your headings are not tagged.
Related
- What PDF remediation costs if you would rather pay someone.
- Section 508 and WCAG requirements to work out what applies to you.