PDF Accessibility Checker

Accessibility problems in a PDF tend to stay invisible until somebody using assistive technology runs into them. This checker opens a document in the browser and reports the signals a machine can actually verify, starting with document setup: title and primary-language metadata, the marked declaration, metadata availability and JavaScript actions.

It then looks at structure signals — per-page structure trees, heading roles, figures, tables and alternative-text attributes — at page-content risks including text presence, image paint operations, image-only page heuristics and document outlines, and at links and forms, covering link text signals, Widget annotations, form-field groups and accessible-label signals exposed by PDF.js.

Results are graded in plain terms. A failure means a rule found a deterministic problem, such as required document metadata that is absent; a warning marks a risk or incomplete signal that deserves investigation but does not determine conformance; a pass means one machine-readable signal was present, not that the whole PDF is accessible. The site is explicit that this is a pre-check, and that decisions requiring meaning, context, visual inspection, keyboard use or assistive-technology testing still need human review. The check runs locally without uploading the document, this release accepts one PDF at a time, and it is free with no account and no payment flow.