PDF/UA Automation
Generate accessible PDFs at scale
Understand the document family first. Then reconstruct, regenerate and validate.
Many organisations publish hundreds or thousands of PDFs that do not meet PDF/UA (ISO 14289). This may look like a technical detail, but it directly affects legal accessibility requirements and practical use with screen readers.
The challenge: inaccessible PDFs
A PDF can look visually perfect and still be inaccessible:
- No tags (Tagged PDF) → screen readers cannot identify headings, paragraphs, lists and tables.
- Incorrect reading order → multi-column layouts, text boxes and footnotes are mixed up.
- Images without alternative text → important information is missing for assistive technology.
- Tables without headers → relationships between row or column headings and cells are unclear.
Manual retagging quickly becomes expensive for large collections. Fully automated repair is not responsible for every existing PDF either. The available source and the structure of the document family determine what can be reproduced reliably.
The solution: a controlled route per document family
We start with one document family and then choose the safest route: regenerate from the source where possible, or reconstruct the existing PDF in a controlled way when source material is unavailable.
- Analysis: inspect source and PDF for text, tags, fonts, links and reading order.
- Grouping: establish which documents demonstrably share the same pattern.
- Regeneration or reconstruction: restore only structures for which enough information is available.
- Validation: check with veraPDF and verify text, links, reading order and visual rendering.
Read more about PDF/UA automation
Results
- A demonstrable boundary: the checked document family and patterns are explicit.
- Repeatable processing: approved rules can be reused within that same family.
- Visible exceptions: documents without enough source information do not silently pass through the pipeline.
Do you have Word or Excel source files? See the separate converter beta for DOCX and experimental XLSX to PDF/UA. It assesses each file individually and supplies a PDF only after successful validation. The reconstruction of existing PDFs described above remains a separate problem and carries no general repair guarantee.
Which document family should be examined first?
Provide one representative PDF and its available source material. The Quickscan makes the structure gap and boundary of the evidence visible.