Accessibility checking starts with structural evidence
Useful automated signals include whether the PDF exposes tagged structure, a document language, a meaningful title, searchable text, bookmarks for longer documents, and labels or tooltips for supported form controls. These clues help prioritize review but do not prove that the semantics are correct.
Important accessibility questions still require human testing
Correct reading order, heading hierarchy, table associations, alternative-text quality, color contrast, keyboard behavior, captions, meaningful link text, and assistive-technology behavior cannot be certified from a small set of parser flags.
- Review the document with a screen reader or other representative assistive technology.
- Check reading order and headings in a dedicated PDF accessibility editor.
- Verify that figures, tables, links, and forms make sense without relying only on visual layout.
Image-only PDFs need an accurate text layer first
A scanned page can look readable while exposing little or no searchable text to assistive technology. OCR can create a text layer, but OCR accuracy and semantic tagging still require review. OpenFileTools does not run OCR in this checker.
Do not treat an automated result as legal or standards certification
A browser report cannot certify WCAG, PDF/UA, ADA, Section 508, legal accessibility, or compatibility with every assistive technology and PDF viewer. Present signals can still be wrong; absent or unknown signals can require deeper inspection.
Use the checker to build a manual review plan
Start with structural findings, fix obvious missing metadata or text-layer problems in an appropriate authoring tool, then perform standards-based and assistive-technology testing on the final PDF. Recheck after edits because remediation can change structure.