Start from the source document
Fixing tags inside a finished PDF is slow, manual work. Fixing the file the PDF came from takes minutes and survives every re-export. Before you export:
- Use real heading styles (Heading 1, Heading 2) instead of bold or large text. These become the H1 and H2 tags a screen reader user navigates by.
- Give every meaningful image alt text in the authoring tool. Mark decorative images as decorative.
- Build tables with the table tool and mark the first row as a header row. Avoid merged cells where you can.
- Use the built-in list tools for lists, not typed dashes or numbers.
- Set the document language, and add a title in the file properties.
Export with tags turned on
Printing to PDF throws the structure away. Use the export or Save As PDF path instead, and check the option that keeps tags:
- Word for Windows: File, Save As, PDF, Options, and tick "Document structure tags for accessibility".
- Word for Mac: File, Save As, PDF, and choose "Best for electronic distribution and accessibility".
- InDesign: export to Adobe PDF (Interactive or Print) with "Create Tagged PDF" checked.
If you only have the PDF and no source, you need a remediation tool that can add tags, such as Adobe Acrobat Pro. Review what automatic tagging produces, because it guesses at headings and reading order.
Set the title, language and bookmarks
Three properties are easy to miss because they do not change how the page looks. The document title should be set and shown in the window title bar (the DisplayDocTitle flag), not the file name. The document language (/Lang) tells a screen reader which pronunciation rules to use. Longer documents need bookmarks so readers can jump between sections.
If the checker reports missing bookmarks, the PDF TOC generator on this site builds an outline from your headings. Title and language are set in the source file's properties or in Acrobat Pro's document properties.
Check it, then check it by hand
Run the checker above. It reports on the text layer, tagging, heading order, figure alt text, table header cells, language, title, bookmarks and encryption, and cites the WCAG or PDF/UA clause for each finding.
Then do what no automated tool can: read the document with a screen reader (NVDA on Windows is free, VoiceOver is built into macOS), confirm the reading order makes sense, and read the alt text to see whether it describes anything. A clean automated result means nothing structural is obviously broken; conformance still depends on that manual review.
What this site can and cannot fix
The checker is free and runs in your browser, so the file is never uploaded. Of the problems it finds, PDFNavigator's tools fix two: missing or broken bookmarks, and owner-password restrictions that block text extraction. Tagging, alt text, table structure, title and language have to be fixed in the source document or a full remediation tool, and each finding names where.
Last updated