Home › Blog › Make a Scanned PDF Accessible
How-toHow to Make a Scanned PDF Accessible (OCR + Remediation)
To a person using a screen reader, a scanned PDF is a blank page. It looks fine on screen, but there is no text underneath — just a picture. Here's the full, honest process for turning one into a genuinely accessible document.
Why scanned PDFs are inaccessible
When you scan a document, you create an image of each page and wrap it in a PDF. There is no text layer — nothing for assistive technology to read, search, or resize. A screen reader reaches the page and announces "graphic", or simply says nothing. The content is, functionally, invisible. Archives, signed contracts, old reports and incoming mail are the usual culprits — and they're covered by the same accessibility laws as everything else you publish.
Step 1 — OCR: create real text
Optical Character Recognition (OCR) reads the image and generates a text layer behind it. This is the unavoidable first step — without text, there is nothing to make accessible. But here's the catch that trips most people up:
Step 2 — Remediate: add the structure
Once real text exists, the document needs the same full PDF/UA treatment as any other PDF:
2a. Reading order
Rebuild the logical order — especially for multi-column layouts and sidebars, which OCR often scrambles.
2b. Tag structure
Mark up headings, paragraphs, lists, tables and links so the document is navigable.
2c. Alt text
Describe stamps, signatures, logos, photos and figures; mark purely decorative marks as artifacts.
2d. Metadata
Set the document title and language so the reader pronounces and announces it correctly.
Step 3 — Validate and prove it
Finally, confirm the result against the standard. Run the document through a formal validator such as PAC 2024, test it with a real screen reader, and keep a signed Accessibility Conformance Report (ACR/VPAT) as evidence for procurement or regulators.
What about poor-quality or handwritten scans?
Usually fixable. Low-quality OCR output is corrected by hand, and handwriting is transcribed manually where the engine can't read it — so the delivered document is complete and accurate, not a best guess. For large archives, this work batches well.
This is precisely our scanned PDF & OCR service: OCR plus full PDF/UA remediation, validated in PAC 2024, with an ACR on every file.
Not sure if your scan has a text layer?
Upload it to the free checker — the score tells you instantly.
Run the free check →