PDF Accessibility CheckerAbout this checker

PDF Accessibility Checker

Check your PDFs

Choose your files, see what needs a closer look, and take the review notes with you.

A first check for accessibility. Reading order, image descriptions and other details still need human review.

Choose your PDFs

Files stay in this browser

Up to 20 PDFs · 20 MiB each · 100 MiB per batch · 200 pages each

Your PDFs stay in this browser. They are not uploaded or saved.

Examples with missing details, tags, a blank page and an unreadable file.

Choose files or try the sample batch.

What this checks

  • Page count, document title and default language
  • Whether tags and page structure can be found
  • Pages where no text can be extracted

Finding tags or text does not show that they are correct. A blank page may need neither.

Continue the review

Reading order, tag quality, image descriptions, forms, links, contrast and text recognition accuracy are not tested here.

For broader checks, use PAC, axesCheck or veraPDF, and inspect the document with assistive technology. Check each tool’s privacy terms before sharing your files.

How to interpret the findings

Mozilla PDF.js reads the document structure. The Marked flag is a declaration; a page structure tree is a presence signal. Neither establishes correct tagging. Text extraction does not confirm text recognition accuracy, visible text, or reading order.

Document language is the catalog’s /Lang value, not a detected language. A title’s presence does not confirm its usefulness or that a viewer displays it. A parser limitation can prevent a property from being retrieved.

Each file has a 60-second time limit. Password-protected files and unsupported XFA forms need a different tool. Limits, cancellations and parser errors remain in the exported report.

Primary references: PDF.js document API, PDF.js page API, and W3C techniques for document titles, language and reading order.