Home › PDF accessibility checker
PDF accessibility checker: nine structural checks in seconds
Upload a PDF. See whether it is tagged, titled, has a language, alt text, headings and a reading order. Free, nothing stored.
Check a PDF
Reads the file's structure tree and metadata only. It does not render pages or extract text, so it cannot judge whether the alt text is good or whether the tags match what is printed. Images and form fields are counted on the first 500 pages. The file is held in memory for the check and discarded.
What the nine checks mean
Every check maps to a PDF/UA (ISO 14289) requirement and to a WCAG 2.2 success criterion via the W3C's PDF techniques. A red row is a definite failure. A green row means the structure exists; a human still has to confirm that it is right.
| Check | What we look for | Why it matters |
|---|---|---|
| Tagged PDF | /MarkInfo /Marked true plus a structure tree (/StructTreeRoot). | Tags are the foundation: without them there are no headings, lists, table cells or alt text for a screen reader to use. PDF/UA 7.1; WCAG 1.3.1. |
| Reading order | At least one element in the structure tree. | The tag tree defines the order content is read. Untagged files are read in drawing order, close to random on a two-column layout. Technique PDF3. |
| Images with alt text | Every Figure tag needs /Alt or /ActualText. A tagged file with page images but no Figure tags gets a "check" (they may all be decorative artifacts); in an untagged file every image is inaccessible. | An image without a text alternative is silent to a screen reader. Technique PDF1; WCAG 1.1.1. |
| Document title | A Title in the document properties (or XMP dc:title), and the viewer preference to display it instead of the file name. | Screen readers announce the title when the file opens. Without it users hear "report-final-v3.pdf". Technique PDF18; WCAG 2.4.2. |
| Document language | A /Lang entry on the document catalog, such as en-US or de. | The language selects the speech voice and pronunciation rules. Technique PDF16; WCAG 3.1.1. |
| Heading structure | H1 to H6 (or generic H) tags in the tree. | Headings are how screen reader users skim and jump between sections. Technique PDF9; WCAG 1.3.1 and 2.4.6. |
| Form fields | Interactive fields (AcroForm) each wrapped in a Form tag. | An untagged field has no name and no place in the reading order. Technique PDF12; WCAG 1.3.1 and 4.1.2. |
| Security | Encryption settings, in particular the permission that allows assistive technology to extract text. | A file that blocks content extraction "for security" also blocks every screen reader. PDF/UA 7.16; WCAG 1.3.1. |
| PDF/UA identifier | pdfuaid:part in the XMP metadata. | Not required by WCAG. It is the file's claim to conform to PDF/UA-1 or PDF/UA-2. |
What PDF/UA and WCAG 2.2 require from a PDF
WCAG applies to PDFs the same way it applies to web pages, and the W3C publishes 23 PDF-specific techniques showing how each success criterion is met inside the format. PDF/UA is the ISO standard for how a PDF has to be built to make that possible: PDF/UA-1 (ISO 14289-1) covers PDF 1.7 files; PDF/UA-2 (ISO 14289-2, published March 2024) covers PDF 2.0 and adds annotations, MathML and richer metadata. The PDF Association's free "Well-Tagged PDF" specification mirrors PDF/UA-2.
In practice a conforming file has a complete tag tree in logical reading order, decorative content marked as artifacts, meaningful images with alternate text, tables with header cells, form fields with tooltips, a title, a language, no restriction that blocks assistive technology, and the PDF/UA identifier. The nine checks above are the machine-decidable subset of that list.
Who has to care: EAA, ADA Title II, Section 508
European Accessibility Act, in force since 28 June 2025
The EAA (Directive 2019/882) has applied to new products and services since 28 June 2025: e-commerce, consumer banking, e-books, transport ticketing, telecom, and the documents that go with them (invoices, statements, contracts, manuals). The technical yardstick is EN 301 549, which points at WCAG. Services that existed before the deadline get a transition period to 28 June 2030; a new PDF published today does not.
ADA Title II: 26 April 2027 and 26 April 2028
The US Department of Justice's 2024 rule under Title II of the ADA requires state and local governments to bring web content, mobile apps and "electronic documents" such as PDFs to WCAG 2.1 Level AA. In April 2026 the DOJ extended the deadlines by a year: entities serving 50,000 people or more must comply by 26 April 2027, smaller entities and special districts by 26 April 2028. PDFs posted before the deadline are exempt only while nobody needs them to apply for or take part in a service; any form in active use has to be fixed.
Section 508 and everyone else
US federal agencies and their suppliers have been bound by Section 508 (WCAG 2.0 AA) since 2018. Private companies are sued under ADA Title III over inaccessible PDFs too: menus, application forms and statements come up in complaints every year.
How to fix each problem
Almost every red row comes from the export step. Fix the source document, re-export with tags on, and most rows turn green together. Repairing a finished PDF in Acrobat is the fallback, not the plan.
Tags and reading order
Word (Windows and Mac): use File, Save As, PDF and, under Options, keep "Document structure tags for accessibility" ticked. Choosing "Minimize size" can silently untick it; check again. Use real heading styles, real lists and the table tools, because the tags come from those styles. Print to PDF never produces tags.
InDesign: map paragraph styles to export tags (Paragraph Style Options, Export Tagging: H1 to H6, P, LI), set the reading order in the Articles panel and tick "Use for tagging order in tagged PDF" in the panel menu, then File, Export, Adobe PDF with "Create Tagged PDF" ticked.
Acrobat Pro: for a finished file with no tags, use Prepare for accessibility, Automatically tag PDF, then open Fix reading order and correct what the autotagger guessed wrong. The Accessibility Checker lists the remaining failures with a Fix option on many of them.
LibreOffice Writer: tick "Universal Accessibility (PDF/UA)" in the PDF export options.
Alt text for images
Add alt text in the source: right-click the image, Edit Alt Text (Word), or Object Export Options, Alt Text (InDesign). Mark purely decorative images as decorative so they become artifacts. In Acrobat, the Accessibility Checker's "Figures alternate text" failure has a Fix action that walks through every figure and lets you mark decorative ones. Keep it short and specific, the same rules as on the web: our alt text checker explains what good alt text looks like.
Document title
Word takes the PDF title from File, Info, Properties, Title. InDesign uses File, File Info, Document Title. In Acrobat: File, Properties, Description, Title, then on the Initial View tab set "Show" to "Document Title" so viewers display it instead of the file name.
Document language
Word writes the proofing language, so set it under Review, Language before exporting. InDesign uses the paragraph style's language. In Acrobat: File, Properties, Advanced, Reading Options, Language. Tag passages in other languages with their own language so the voice switches.
Headings
Use Heading 1 to Heading 6 styles in Word, or map styles to H1 to H6 in InDesign, and do not skip levels. Bold, larger text is not a heading to a screen reader. In Acrobat, the Reading Order tool lets you mark a block as Heading 1 to 6.
Form fields
Fields created with Acrobat's Prepare Form tool are tagged automatically when the document already has tags; add a Tooltip to every field (Properties, General), because the tooltip is the name a screen reader announces. Fields added to an untagged document have no structure: tag the document first, then run Prepare Form again.
Security and permissions
If you password-protect a PDF, keep "Enable text access for screen reader devices for the visually impaired" ticked in Acrobat's Document Security settings. Turning it off blocks every screen reader, not just copy and paste. A file that needs a password to open cannot be checked here at all; remove the open password, or check the unprotected original.
What this tool cannot tell you
It reads structure, not meaning. It can prove a figure has alt text, not that the alt text says something useful; that headings exist, not that they are on the right text. It does not check tables, lists, links, colour contrast or scanned pages without a text layer. For those, run Acrobat's full checker or a PDF/UA validator after this one, and put the document in front of a screen reader user before it goes out.
The website that links to your PDFs is the other half. The free AdaClear scan checks a page in a real browser with the full axe-core rule set, and the paid plans watch the whole site so a template change that breaks it is caught the same day.
PDFs done. Now the site that hosts them.
Free 30-second scan: score, every WCAG failure automated testing can find, and the fixes in plain language. No overlay, no signup.
Scan my site free →Frequently asked questions
What does "tagged PDF" mean?
A tagged PDF carries a hidden structure tree that tells assistive technology what each piece of content is (heading, paragraph, list, table cell, figure) and in what order to read it. Without tags a screen reader has to guess from where text sits on the page, and images have no text alternative at all.
Is my PDF stored when I upload it?
No. The file is held in memory only for the seconds the check takes, is never written to disk, and is discarded as soon as the result is sent back. The file name is not logged either.
Does passing every check make the PDF compliant?
No. These are the structural checks a machine can decide with certainty. Whether the tags match the visible content, whether the alt text is meaningful and whether tables are tagged correctly still need a human, or a full PDF/UA checker such as PAC. A file that fails these checks is definitely not accessible; one that passes is ready for that deeper review.
Why does it say my images have no alt text when I added some in Word?
Usually because the export dropped the tags: "Print to PDF", "Minimize size" with the tags option unticked, or a third-party converter. Re-export with "Document structure tags for accessibility" ticked and check again.
My PDF is a scan. Why does it pass so little?
A scanned PDF is a picture of a page: no text, no tags, no headings, so a screen reader reads nothing. Run OCR (Acrobat: Scan and OCR, Recognize Text), tag the result, then check again. Better, find the original file and export from that.
Which tool should I check a PDF with before publishing?
This one first, because it is instant and catches the structural failures that make everything else moot. Then Acrobat Pro's Accessibility Checker or a PDF/UA validator (PAC is free) for tables, lists and tag correctness, and a pass with a screen reader such as NVDA.
Sources
- W3C, PDF Techniques for WCAG: PDF1, PDF9, full list (PDF3, PDF12, PDF16, PDF18).
- PDF Association, ISO 14289-2 (PDF/UA-2) has arrived and the ISO 14289-1 (PDF/UA-1) resource page.
- ISO, ISO 14289-2:2024.
- EUR-Lex, Directive (EU) 2019/882, the European Accessibility Act.
- US Department of Justice, Fact sheet: ADA Title II web rule (dates as extended April 2026).
- Federal Register, Extension of compliance dates, 20 April 2026.
- Adobe, Create and verify PDF accessibility (Acrobat Pro) and Use tags for accessible PDFs in InDesign.
- Microsoft, Create accessible PDFs.