Accessible PDF: What It Means in Practice

Updated · 4 min read

On this page
  1. What does an accessible PDF actually contain?
  2. Accessible PDF versus regular PDF, in concrete terms
  3. What does a screen reader do with each?
  4. A worked example: reading order before and after
  5. How do I know if a PDF is already accessible?
  6. How do you get from one to the other?

An accessible PDF is a PDF that a screen reader can read in the right order and understand: it carries a tag structure that names each element, a reading order that matches the page, and the language, title and image alternatives that assistive technology needs. A regular PDF may look identical on screen and be completely unreadable to someone who cannot see it, because it is only a picture of a layout.

What does an accessible PDF actually contain?

The requirements are concrete, and each one maps to a real behaviour:

  • A tag structure. A tree that labels each element: H1 for the title, H2 and H3 for sections, P for paragraphs, L and LI for lists, Table with header cells, Figure for images.
  • A reading order. The order in which the screen reader announces content, which must follow the logical flow rather than the order objects happen to sit in the file.
  • Alt text on meaningful images, and decorative images marked as artifacts.
  • A document language (/Lang) so the screen reader uses the right pronunciation.
  • A document title, plus the flag that tells the reader to use it, so the tab shows a real name rather than a filename.
  • Contrast that meets the minimum ratio and colour that is never the only way information is conveyed.
  • Labels on any interactive form fields.

That list is essentially the PDF side of WCAG 2.1 AA (the WCAG 2.1 specification) and of PDF/UA, or ISO 14289. A file that satisfies the list satisfies both.

Accessible PDF versus regular PDF, in concrete terms

Regular PDFAccessible PDF
What it isText and images at coordinatesTagged elements in a structure tree
Reading orderWhatever order objects sit inThe logical reading flow
HeadingsBold, larger textH1-H6 tags, so navigation works
ImagesPixels on the pageFigures with alt text, or artifacts
LanguageNot declared/Lang set, e.g. en-US
TitleThe filenameA real document title
Screen readerGuesses from layoutFollows the tags

The visible page can be identical in both columns. Everything that differs lives behind the page, in the structure and metadata.

What does a screen reader do with each?

On a tagged PDF, a screen reader announces the document title, then reads the content in the tagged order. It can jump between headings, list the headings as an outline, read a table by cell with its header announced, and skip decorative images. A user can navigate a 40-page report by heading the way a sighted reader skims it.

On an untagged PDF, the screen reader falls back to guessing. It reads text in the order the objects sit in the file, which in a two-column newsletter often means reading the left column and the right column interleaved, line by line, producing nonsense. Headings are just slightly larger text, so heading navigation does nothing. Images are announced as “image” with no description, or skipped. A user can still listen to the whole document, but they cannot navigate it, and they may hear content in the wrong order.

The screen reader PDF guide walks through what a real user hears on each kind of file.

A worked example: reading order before and after

A two-column program for a public meeting. On the page, the left column holds the agenda and the right column holds speaker notes.

Untagged, what the screen reader reads:

“Agenda item one. Notes for item one. Agenda item two. Notes for item two. Agenda item three…”

Tagged, what the screen reader reads:

“Agenda. Item one. Item two. Item three. Speaker notes. Notes for item one. Notes for item two…”

The second reads correctly because the tags say which text is a heading, which is body, and which column comes first. The visible page never changed.

How do I know if a PDF is already accessible?

You cannot tell by looking, because everything that differs sits behind the page. Two quick signals help:

  • Try to select a heading in the file. If you can highlight the text, it has a text layer; if you cannot, it is a scan and needs OCR before anything else.
  • Run it through a free checker such as PAC 2024 or veraPDF. If “Tagged PDF” fails, the file has no usable structure yet.

Neither signal tells you whether the alt text is meaningful or the reading order makes sense; those need a screen reader. The PDF accessibility checker guide covers the tools, and PDF accessibility testing covers the human half of the job.

How do you get from one to the other?

If you have the source, export it with tags on, which is the reliable route and is covered in how to make a PDF accessible. If you only have the PDF, you remediate it: by hand in Acrobat Pro, or with a hosted tool. Either way, verify the result against the accessible PDF checklist before publishing.

A remediation aid such as TagRight writes the tag structure, reading order, language and title into a PDF you did not author, leaves the pages looking the same, and returns a list of what still needs a human, such as alt text for images. It is not a certification, and a high-stakes file should be confirmed with PAC 2024 or veraPDF. The free tier is on the pricing page, and you can tag a PDF free to see the difference on your own file.

Frequently asked questions

What is the difference between a PDF and an accessible PDF?

A regular PDF is a picture of a layout: text and images placed at coordinates with no indication of what each piece is. An accessible PDF adds a tag structure that names each element, sets a reading order, and records the language, title and image alternatives a screen reader needs.

Does an accessible PDF look different?

It should not. Good remediation adds the tags and metadata behind the page while leaving the visible layout unchanged. If a tool changes how your document looks, treat that as a problem to check.

Is every PDF that opens in a screen reader accessible?

No. A screen reader will open any PDF, but on an untagged file it has to guess the structure from the visual layout, which often produces the wrong reading order. Opening is not the same as being readable.

Do accessible PDFs need alt text on every image?

Only images that carry meaning need alt text. Decorative images should be marked as artifacts so they are skipped. An image that conveys information and has no alternative leaves that information unavailable.

What is the single most important part of an accessible PDF?

The tag structure. Without it there is no reading order, no heading navigation and nowhere to store alt text. Everything else in an accessible PDF depends on tags being present and correct.

Tag a PDF free