How does this checker audit PDF/UA-1 and WCAG 2.2 rules?
Inside a browser Web Worker, the parser reads your file’s internal dictionary objects, catalog flags, and page content streams without uploading any document bytes. It compares those PDF structures against eleven machine-verifiable requirements drawn from ISO 14289-1 and W3C accessibility guidelines.
Inspecting the structure tag tree and MarkInfo
When you run an accessible PDF test, this PDF accessibility checker scans the document catalog for a populated /StructTreeRoot entry and /MarkInfo with /Marked set to true. Under the PDF Association ISO 14289 (PDF/UA) overview, custom tags must map back to standard semantic roles via /RoleMap.
Our walker follows /RoleMap chains up to 10 hops deep so vendor tags resolve cleanly to P or H1, inspecting up to 15,000 nodes at depth 64. Catalog /Lang values match BCP-47 tags like en-US, while /Suspects true raises a warning.
If your publishing pipeline produces digital reflowable books alongside fixed PDFs, pair structural checks across formats locally. You can validate EPUB 3 packages and EAA tags with the same zero-upload privacy model.
Scoring the 11 automated accessibility checkpoints
Every checkpoint evaluates to Pass, Warning, Fail, or N/A across five filterable result categories. Applicable checks determine the 0 to 100 Score, giving 1.0 for Pass and 0.5 for Warning while excluding N/A items so unrelated rules never distort your rating.
Eleven automated checks across structure, metadata, and forms
Eleven automated checks map directly to PDF dictionary keys and W3C WCAG 2.2 Recommendation criteria. Filter failures across Structure, Metadata, Images, Navigation, or Forms tabs.
| Checkpoint | Category tab | Standard clause | Passing condition verified in PDF |
|---|---|---|---|
| Tagged logical structure | Structure | PDF/UA 7.1, WCAG 1.3.1 | Populated /StructTreeRoot and /MarkInfo /Marked true |
| Heading hierarchy | Structure | PDF/UA 7.4, WCAG 2.4.6 | Starts at H1 with zero skipped levels across H1–H6 |
| Data table headers | Structure | PDF/UA 7.5, WCAG 1.3.1 | Every PDF Table subtree contains at least one TH cell |
| Title and window flag | Metadata | PDF/UA 7.1, WCAG 2.4.2 | Non-filename /Title or XMP dc:title plus /DisplayDocTitle |
| Primary language | Metadata | PDF/UA 7.2, WCAG 3.1.1 | Valid BCP-47 tag inside catalog /Lang |
| Text vs scanned images | Images | PDF/UA 7.1, WCAG 1.4.5 | Pages include real text operators (Tj or TJ), not scans |
| Figure alternate text | Images | PDF/UA 7.3, WCAG 1.1.1 | Every <Figure> and <Formula> tag holds /Alt or /ActualText |
| Interactive form tooltips | Forms | PDF/UA 7.18, WCAG 4.1.2 | Every AcroForm field and Widget provides a /TU label |
| Annotation tab order | Navigation | PDF/UA 7.18.3, WCAG 2.4.3 | Annotated pages set /Tabs /S for structure order |
| Outline bookmarks | Navigation | WCAG 2.4.5 | Documents with 5 or more pages include catalog /Outlines |
The validator checks Info /Title and XMP dc:title, rejecting placeholders or .pdf filenames. Form audits inspect /AcroForm /Fields and Widget annotations for /TU tooltips.
Static contracts with unlabelled form fields do not always need editable boxes. You can prevent interactive widget drift before static release so entries print cleanly as text.
Inspecting scanned pages, alt text, and the tag tree viewer
Content-stream scanning catches image-only scanned PDFs missing text layers even when a document claims to be tagged. PageStreamScanner inspects each page’s /Contents array and Form XObjects for Tj or TJ text operators versus Image XObjects.
Detecting image-only scans and missing figure alt text
When pages draw bitmaps without text operators, the Selectable text vs image-only scanned pages card offers Run OCR to add searchable text. Adding OCR text creates selectable characters before you tag reading order.
For text PDFs, Alternative text on Figure elements flags <Figure> and <Formula> nodes missing /Alt or /ActualText. Under the W3C WCAG 2.2 PDF1 alternative text technique, meaningful graphics require descriptions while decorative rules become artifacts.
Tip: Click Show tag tree after fixing missing alt text to verify decorative borders no longer appear as empty
<Figure>nodes.
Reading the structure tag outline and table headers
Show tag tree renders up to 120 structure elements with tag names, page numbers (p.1), and [Alt: ...] attributes so you can audit logical sequence in the PDF tag tree viewer. Data table headers checks PDF Table subtrees up to depth 20 for TH, /Scope, and /Headers. Because parsers ignore raster colors, check embedded charts with our WCAG 2.2 and APCA contrast ratio calculator.
Automated PDF accessibility audit vs manual review for EAA
Software checks alone cannot replace human verification, because rule engines catch missing /Lang strings, unset /DisplayDocTitle flags, and untagged pages while people confirm reading logic and contrast.
Meeting European Accessibility Act and ADA timelines
Under the European Accessibility Act (Directive 2019/882), binding accessibility rules took effect on June 28, 2025 for consumer publications, statements, and public services. In the United States, ADA Title II likewise cites WCAG 2.1/2.2 Level AA benchmarks.
- Automated scope: Verifies all 11 catalog, tag tree, tooltip, and scanned-page checks locally.
- Manual review scope: Confirms accurate figure alt text, sidebar reading order, and visual color clarity.
When auditing charts or color-coded status tables, never rely on hue alone. Inspect screenshots in a Color blindness simulator so series remain clear across color-vision deficiencies.
Exporting CSV logs and A4 PDF compliance reports
Click Export CSV for a UTF-8 BOM spreadsheet (-accessibility-audit.csv) protected against formula injection. Each row logs the score, clause reference, check status, summary, and failing page indices.
Click Download PDF report to export a formatted A4 (595 × 842 pt) compliance sheet (-accessibility-report.pdf). Both generators run inside tab memory without uploads, keeping sensitive legal, medical, and student records private.