PDF accessibility checker
See what a PDF is missing before a screen reader user finds it. The checker reads the file in your browser and reports its tags, language, title, fonts, alt text, links, form fields, bookmarks and encryption โ with the fix beside every failure.
What does this PDF accessibility checker do?
It opens a PDF in your browser and reports the accessibility signals that live in the file itself: whether the document is tagged, whether it carries a document language, a title and a structure tree, whether its fonts can be read aloud, whether images are likely to be missing alt text, whether links and form fields are labelled, and whether the file is encrypted. Each check passes, warns or fails, and the results add up to a score. The PDF never leaves your device.
Your PDF never leaves this device. The file is parsed by your own browser, including any compressed streams it contains. No document is uploaded, no third party sees it, and nothing is stored on a server. Closing the tab discards everything.
Empty state. No PDF checked yetAdd a file above and its accessibility report will appear here.
Not checked
A structural scan, not a conformance audit. Reading order, writing quality and whether alt text is meaningful still need a human with a screen reader.
File details
Checks
What gets checked
- Tagged document. The Marked flag and a structure tree, which give a screen reader headings, lists and a reading order.
- Document language. The language a screen reader uses to choose pronunciation rules.
- Document title. The title announced instead of the file name, and whether the viewer is asked to show it.
- Readable fonts. Fonts with a ToUnicode map, without which copied and spoken text can turn into gibberish.
- Images have alt text. Alternative text entries in the file, or the absence of images entirely.
- Labelled links. Links and annotations that sit inside the structure tree so they can be reached and announced.
- Labelled form fields. A tooltip or accessible name on every field, so an unnamed edit box does not read as blank.
- Selectable text. Real text-drawing commands rather than a picture of text, which is what a scan produces.
- Bookmarks and metadata. An outline for long documents, XMP metadata, and a PDF/UA identifier when one is claimed.
- Encryption and XFA. Protection that may block accessibility tools, and XFA forms that most readers cannot handle.
What it cannot tell you
A file can pass every structural check and still be unusable. This tool cannot judge whether the reading order makes sense, whether an image description is accurate, whether a table is marked up correctly, or whether the language is plain enough. Use it as a fast first pass, then test the document with a screen reader.
How to check a PDF for accessibility
Choose the PDF
Drop a PDF onto the area above or use the file picker. One file at a time, up to 50 MB. The file is read on your own device and is never uploaded.
Read the score and the checks
The checker reports the document tags, structure tree, language, title, fonts, images, links, form fields, bookmarks and encryption. Each check passes, warns or fails, and the score summarises the result.
Fix the failures at the source
Every warning and failure names its fix. Most are fixed in the application the PDF was authored in โ enable tagging on export, set the title and language, add alt text and accessible names โ then export again and re-check the file.
What a tagged PDF means
A PDF describes where marks go on a page, not what those marks mean. Tagging adds the meaning back: this run of text is a heading, this is a list, this is a table row, read this before that. Without tags a screen reader has to guess the reading order from the position of the text on the page, and on a two-column layout or a sidebar it guesses wrong.
The checker looks for the two things that make a PDF tagged โ the Marked flag and a structure tree โ and then for the properties that hang off them: a language on the document, a title, alternative text on images, accessible names on form fields and annotations inside the structure tree.
Why a high score is not a certificate
Accessibility is a property of the reading experience, not of the file. A structural scan can prove that alt text is absent; it cannot prove that present alt text is useful. That is why the report separates passes from warnings, and why the disclaimer stays on the page rather than in a footer: a green score is a good sign, not a conformance claim.
If the PDF turns out to be a scan, the text has to be recovered before anything else: the screenshot-to-text tool reads an image of text in your browser, and the text fixer cleans up what comes out. For the document's words rather than its structure, use the text tools.
Common questions
Is my PDF uploaded anywhere?
No. The file is read by your own browser, and no request carrying it is ever made. PDFs are often contracts, reports and forms, so this is the one thing the tool has to be right about: closing the tab discards everything.
What does the checker actually check?
The document-level properties a screen reader depends on: a Marked/StructTreeRoot tag tree, a document language, a document title, book marks for navigation, fonts with ToUnicode maps, images without alt text, links and form fields without accessible names, XFA forms and encryption. Every check shows what was found and how to fix it.
What is a tagged PDF?
A tagged PDF is one whose structure is written into the file: headings are marked as headings, lists as lists, tables as tables, and the reading order is defined rather than guessed from where the text sits on the page. It is what makes a PDF readable by a screen reader, searchable in a real way and re-flowable. A PDF exported straight from a scanner or a print driver usually has no tags at all.
Does a high score mean the PDF is WCAG conformant?
No, and this tool will not pretend otherwise. A score summarises the structural checks that can be made by reading the file itself. It cannot judge writing level, whether alt text says the right thing, or the reading order of a complex layout โ those need a human with a screen reader. Treat the score as a fast list of what is obviously missing, not as a conformance claim.
How do I fix the problems this checker finds?
In the tool the PDF was made in. Enable automatic tagging on export, set the document language and title in the document properties, add alt text to every meaningful image, give links and form fields accessible names, and re-export. The check list on this page names the fix for each individual failure.