paridoc by eFile

Making scanned PDFs accessible.

A scanned PDF is a picture of a page: there's no text in it for a screen reader to read. Paridoc reads each scan with OCR, then rebuilds it as structured, accessible HTML with headings, tables and reading order, checked against WCAG 2.2 AA.

More common than you'd think

Signed contracts, letters from the public, old ordinances, recorded documents, anything that went through a copier on its way to the website. Of the PDF pages Paridoc has converted so far, about 30% were scans. On a scanned page, a screen reader user hears nothing at all, or just "image".

Why OCR alone isn't enough

OCR, optical character recognition, turns the picture of text into text. That's the first step, not the last. Raw OCR output has no headings, tables come out as rows of loose numbers, and two-column pages are often read straight across both columns. WCAG requires structure as well as text.

What Paridoc does with a scan

  1. Recognizes the text on each page with OCR.
  2. Reads the layout: columns, headings, tables, captions, headers and footers.
  3. Rebuilds the page as HTML, with headings in order, tables with header cells, and text in the order it's meant to be read.
  4. Describes what isn't text: signatures, stamps, seals, photographs and charts.
  5. Checks the result in a real browser against WCAG 2.2 AA with axe-core.

If the OCR misread something, flag it and our team corrects it against the original, with every fix recorded. How it works.

Questions

Why can't a screen reader read a scanned PDF?

A scanned PDF is an image of each page. There is no text in the file, so a screen reader has nothing to read, and search, copy and zoom-to-reflow don't work either.

Is running OCR enough to make a scan accessible?

No. OCR turns the image into text, but the result has no headings, no table structure and often the wrong reading order. It still needs to be structured before it meets WCAG.

What happens to signatures and stamps?

They are kept as images in the accessible version and given a written description, so a screen reader user knows a document was signed or stamped.

Updated September 2026

Send us your website's address and we'll count your PDFs and send you a price: paridoc@efile.systems