Making a PDF Accessible: What Actually Helps
An inaccessible PDF is invisible to a screen reader: the user hears "image" and nothing else. Most of the fix is about making the document machine-readable, and the single biggest step is the one people skip.
The one thing that matters most
The document must contain real, selectable text. If it is a scan, a screen reader has nothing to read — regardless of anything else you do.
Running OCR adds a text layer, which turns an unreadable scan into a readable document. That single step does more than every other accessibility measure combined.
What else helps
- Real headings, not just large bold text. Tags let a screen reader announce "heading level 2" and let the user jump between sections.
- A document title in the metadata. Screen readers announce it when the file opens.
- A logical reading order. Multi-column layouts without tags are read straight across the page, producing nonsense.
- Alt text on images that carry meaning. Decorative images can be marked as such.
- Sufficient contrast. Grey text on white fails for low vision, and it fails in print too.
- Bookmarks for anything over about ten pages.
What is mostly theatre
Adding a "this document is accessible" statement without doing the work above helps nobody.
Converting an accessible Word document to PDF badly can destroy the tagging. If accessibility matters, export with the tagged-PDF option and verify.
A realistic check
- Select all the text. If nothing selects, it is a scan — run OCR.
- Open the bookmarks panel. Are there entries?
- Open the document properties. Is there a title?
- Try navigating with the keyboard only — Tab should reach links and fields.
- If it passes those four, it is better than most PDFs in circulation.
Privacy
Accessibility work often happens on public-sector and health documents, which carry personal data. OCR and conversion through an online service uploads the whole file.
A browser-based tool keeps the document on your device through OCR and repair.