PDF accessibility: why PDFs fail screen readers, and how to fix them

The short answer: a screen reader can only read a PDF properly when the PDF has tags, a hidden structure that says what is a heading, a paragraph, a list, a table or a picture, and in what order to read them. Many PDFs have no tags, and scanned PDFs have no text at all. The fix is a rebuilt, tagged PDF.

What a screen reader needs

People who are blind or have low vision often use a screen reader, software that reads a page aloud. In a PDF it follows the tags. Without them it reads the words in whatever order they happen to be stored, which can jumble columns and tables, or it finds nothing at all. The WCAG 2.1 AA standard in the federal web rules covers documents such as PDFs that are posted on a website.

Three common problems

  • No tags. The document has text, but no structure to read it by.
  • Scanned pages. A scanned PDF is a picture of paper. It has no real text, so there is nothing to read.
  • No title or language. A screen reader announces the title, and needs the language to pronounce words correctly.

How to check your own PDFs

In a PDF reader's document properties, look for whether the document is tagged, its title and its language. Try selecting text: if you cannot, the page is probably a scanned image. Our free scan checks up to five PDFs linked from your website for tags, real text versus page images, title and language.

How EveryRamp rebuilds a PDF

  1. It reads the document's text and layout, including images of any scanned pages.
  2. It rebuilds the structure: lines into columns and paragraphs, headings by their size, lists by their bullets or numbers, simple tables by their aligned columns. A header, footer or page number that repeats on every page is kept once. Pictures keep their place and get a written description. Scanned pages are read into real text, and a page is only transcribed when two separate AI models agree on its text.
  3. It produces a new tagged PDF with a title, a language and bookmarks.

Honest limits

The rebuilt PDF is a new document with its text reflowed, so it can have a different number of pages from the original. Complex forms and complex tables may still need a person to review them. Descriptions written automatically should be read once before publishing. Your original file stays yours to keep or replace. We rebuild a document up to 1,000 pages and 25 MB; a larger document is still tested and reported on, but is not rebuilt.

EveryRamp tests websites and PDFs against WCAG 2.1 AA and writes the fixes. You can start with a free scan of your website, see a sample report, or compare the packages.

This page is general information, not legal advice.

Guides