Page and tag evidence
Use validator rules, structure elements, object references, and MCIDs when available.
The document accessibility API creates one preflight boundary in front of your CMS, DAM, PIM, or generation service. It routes PDF and DOCX to format-native workers while returning one decision and evidence model.
For content platforms that cannot maintain separate Word and PDF checking stacks
Illustrativeapplication/pdfPDF/UA + tag/object evidenceOOXML packageStyles + parts + relationshipsThe platform can keep its own document state machine while AccessPreflight handles upload validation, asynchronous workers, and evidence.
Create an upload with media type, size, SHA-256, and source-retention choice.
Run PDF or OOXML-native layers under the selected versioned profiles.
Allow, warn, block, or route review with finding and report links.
Normalization aligns workflow fields and outcomes. It does not pretend that a PDF object and a Word paragraph are the same locator.
Use validator rules, structure elements, object references, and MCIDs when available.
Use package parts, paragraphs, styles, relationships, tables, and drawings.
Use stable identity, severity, confidence, coverage, repair, and gate fields.
PDF and DOCX inputs can share a workflow while using format-appropriate technical profiles. The returned coverage shows which rules ran for each file.
The normalized delta stays useful to publishing systems while PDF and DOCX locators remain attached to the changed finding.
Illustrative publishing state: review required
Upload a PDF or DOCX in a free project, inspect its native evidence, and plan the API boundary around the document system you already operate.
Automated preflight is not an accessibility certification or legal determination.