← Back to the PDF Accessibility Tag Checker
Several free PDF/UA and Section 508 checkers already exist online -- PDFix, PDF Audit, freepdfchecker.com, and axesCheck among them -- and there is a well-known free desktop tool, PAC. The question worth asking before choosing one is not which finds the most issues; the structural checks in this class of tool are largely the same checks. It is what happens to the file you hand it, and whether you can check more than one at a time.
| Where the file goes | The surveyed free online checkers (PDFix, PDF Audit, freepdfchecker.com, axesCheck) read your
PDF by first uploading it to their server. This checker's browser page reads the file with the
browser's own File and DecompressionStream APIs, entirely in your tab --
nothing is sent anywhere. For a government form, a permit, meeting minutes, or an HR document, that
is often the reason to be reluctant to hand a file to an unfamiliar site in the first place. |
|---|---|
| One file or a folder | PAC, the classic free desktop tool, checks one file at a time. None of the surveyed free online
checkers offer a batch or folder mode either. This tool's CLI, cli/check.mjs, walks a whole
folder recursively (node cli/check.mjs ./document-library/ --json) and exits non-zero if any
file has a finding -- so a document library of hundreds of PDFs is one command, and the same command
can gate a CI job on newly published documents. |
| What it costs to run again | Re-checking a PDF after a fix with an online checker means uploading it again. Re-checking here means dropping the file again or re-running the CLI -- no account, no queue, no per-file limit. |
The structural checks themselves are the ordinary PDF/UA and Section 508 building blocks any tool in this class looks at: a tagged structure tree, a document title, a declared language, image alt text, an extractable text layer instead of an un-OCR'd scan, and accessible names on form fields. None of that is unique to this tool, and none of it is a substitute for a manual or assistive-technology review -- see the checker's own disclaimer on index.html for exactly what a clean report does and does not mean.
The ones surveyed here do -- PDFix, PDF Audit, freepdfchecker.com, and axesCheck all ask you to choose or drop a file that is then sent to their server to be scanned. This checker reads the file with the browser's own File and DecompressionStream APIs entirely in your tab, so a government form, a permit, or an HR document never has to leave your machine to be checked.
Yes, with the CLI: node cli/check.mjs ./document-library/ --json walks a folder recursively and exits non-zero if any file has a finding, so it can gate a CI job or a scheduled task. The classic free desktop tool, PAC, checks one file at a time with no batch mode, and none of the surveyed free online checkers offer a batch or folder option either.
No. This is evidence, not a determination -- specific, automatable structural checks (tagging, title, language, alt-text coverage, text layer, form labels), not a full PDF/UA or Section 508 conformance audit and not a substitute for a manual or assistive-technology review. A clean report means these checks found nothing, not that the document is compliant.
It reads the file's internal structure rather than how it renders: whether the document catalog declares /MarkInfo /Marked true and a /StructTreeRoot, a non-empty /Title and a well-formed /Lang tag, whether image elements in the structure tree carry non-empty /Alt text, whether a page has an extractable text layer or is likely an un-OCR'd scan, and whether interactive form fields have an accessible name. None of that is visible just by looking at the rendered page.
← Back to the checker to drop in your own PDF, or use the CLI to batch-scan a folder.