Make scanned documents more accessible by checking reading order, headings, tables, language, alt text, and OCR accuracy.

OCR is the first accessibility step

OCR can expose text trapped in a scan, but an accessible document also needs meaningful order, semantic headings, tagged structures, language settings, and accurate alternatives.

  • Recognizable text
  • Logical reading order
  • Real heading hierarchy
  • Tagged lists and tables
  • Useful image alternatives

What OCR fixes

OCR can expose words that were previously trapped in page images. This enables search, selection, and a starting point for assistive technology. It does not automatically add correct headings, lists, table headers, or alternative text.

Recognition provides

Characters that can become searchable, selectable, and available to assistive technology.

Remediation provides

Relationships and meaning such as headings, lists, tables, labels, order, and alternatives.

When interface text cannot be selected, the Screenshot to Text tool can provide a draft for review and remediation.

Reading order

Multi-column pages, sidebars, footnotes, and forms can be announced in an incoherent sequence. Test the logical order independently from the visual page. A clean-looking PDF may still produce unusable speech output.

  1. Start at the title

    Follow the content without relying on its visual position.

  2. Move through each region

    Check columns, sidebars, captions, footnotes, and form labels in sequence.

  3. Listen for changed meaning

    Stop when an out-of-order passage separates a label from its value or a note from its subject.

Start by inspecting the extracted sequence with the scanned PDF to text tool, then remediate the document structure separately.

Language and structure

Set the document language, tag headings by hierarchy, identify lists and tables, and give informative images appropriate alternatives. Decorative images should not create noise.

Language
Set the document language so pronunciation and character interpretation have the right context.
Headings
Use a logical hierarchy that reflects sections rather than visual size alone.
Tables
Identify header cells and relationships instead of flattening every cell into a text stream.
Images
Describe informative images and keep decorative images from adding noise.

For lossless interface captures, the PNG to Text tool preserves the source pixels while exposing OCR diagnostics.

Quality assurance

Search for known phrases, copy representative passages, navigate by headings, and test with a screen reader. Review page numbers, hyphenated line endings, and characters that OCR commonly confuses.

Text check

Search known phrases and compare representative copied passages with the scan.

Structure check

Navigate by headings and verify tables, lists, and form labels.

Use check

Test the final reading sequence with assistive technology, not visual appearance alone.

The searchable PDF versus plain text guide explains what each output preserves and loses.

Method and sources

This guide distinguishes text recognition from document remediation. Accessibility claims require separate checks for reading order, headings, tables, language, and assistive-technology behavior. Reproduce the technical behavior with the documented Free-OCR.net OCR methodology.