Taggart
Hebrew PDF remediation software
Software that finds and repairs what breaks Hebrew PDFs — text stored backwards, the wrong language, columns read left to right — leaving the page as it was.
The failure the standard does not test
Many Hebrew PDFs — from report generators, print drivers, legal and government systems, and some OCR engines — store their text in visual order: each line reversed in advance and painted left to right. The page looks perfect. Copy a line and it pastes backwards; search finds nothing; a screen reader reads it from the end.
PDF/UA and the Matterhorn Protocol check tags, structure, language, fonts and alternate text. None of their conditions asks in which direction the characters were stored. So these files pass veraPDF, PAC and Acrobat’s checker and are still unreadable. The failure, in detail →
Taggart was built around that gap.
What it does with a Hebrew document
Finds visual order, line by line, from two independent signals: the direction the producer painted each line in, and where Hebrew’s final letters — ך ם ן ף ץ, which only ever end a word — fall. For every line it shows the text as stored, the text in logical order, and the evidence. Lines where the evidence is weak are reported as undecided and shown to you, never counted as fine.
Repairs it without touching the page. The logical text goes into ActualText with the line’s language. When a producer puts several lines of a paragraph into one piece of content — common in real documents — every line gets its own text and none is left silent. Every repaired page is re-rendered and compared with the original at structural similarity ≥ 0.999; a repair that changes the picture is discarded.
Sets the language where it is missing, per run, so Hebrew is read in a Hebrew voice. Why the language matters →
Proposes a right-to-left reading order — right column first — for pages whose order runs the wrong way, shown to you as a change you accept rather than applied silently. Reading order →
Reads scanned Hebrew with OCR on the scanned pages only, and checks the result for direction like everything else.
And everything else a PDF needs: tagging, headings, lists, tables, titles, links, fonts — checked against PDF/UA-1, WCAG 2.1 AA and IS 5568, with a report that names every failure by its Matterhorn condition.
What it does not do
It does not write alternate text: it shows each image with what the document says about it and you write the description. It does not decide the conditions that need a person, and it will not write a conformance claim while one is open. And it does not make a legal determination — whether a document meets IS 5568 is an accessibility certifier’s call.
On a real document
On a real four-page Hebrew document stored in visual order, one glyph to a text operation, the repair took it from 135 findings to one — the PDF/UA identifier, which is written only once a person confirms the judgement calls — with every page unchanged. The synthetic test corpus had never reproduced that producer’s quirks; every serious Hebrew defect this product has fixed was found on real files.
Try it on yours
Check a Hebrew file free — no account, and the right-to-left tab shows every line. The Hebrew guides go further: Hebrew PDF accessibility, text that copies out reversed, and in Hebrew.
Frequently asked questions
What goes wrong with Hebrew PDFs that does not go wrong with English ones?
Three things, all invisible on the page. Text stored in visual order — pre-reversed — so copy, search and screen readers get it backwards. A document language left as English, so a screen reader reads Hebrew in an English voice. And columns read left to right, so a two-column page reads as interleaved halves. None of the three is tested by PDF/UA, which is why a file can pass veraPDF, PAC and Acrobat and still be unusable in Hebrew.
Does the repair change how the document looks?
No. The logical-order text is attached as ActualText and the language as /Lang; the painted glyphs do not move. Every repaired page is rendered and compared with the original, and a repair that changes the picture is thrown away rather than returned.
Does it decide things for me?
Only the mechanical things. Lines where the evidence for visual order is weak are reported as undecided and shown to you, not counted as fine. A reading order, a heading, a table header or a description is a proposal you accept, and the report records who decided what and when.
Is the interface in Hebrew?
Yes — the application, the reports' rule titles for the conditions Hebrew documents actually meet, and the right-to-left panel, which shows every line as stored, in logical order, and the evidence for the verdict.
What does it cost?
Checking is free and needs no account up to 60 pages. With a free account the repair runs on your own file, up to 60 pages, so you see the result. A plan, $29 a month, is for downloading the repaired file and its report.
Related guides
- Hebrew PDF accessibility: the three failures that matterVisual-order storage, wrong document language, and left-to-right column order. All three are invisible to a sighted reviewer and to every mainstream checker.
- Hebrew or Arabic text reversed when copied from a PDFYou copy a line of Hebrew or Arabic out of a PDF and it pastes backwards. Why it happens, a ten-second test that tells file from app, and how to fix it.
- ActualText in PDF: what it is and when to use itActualText changes what a PDF says without changing how it looks. When it is the right fix, how it differs from Alt, and the mistakes that silence text.
Run your own file through it. The validator is free and unlimited — no page cap, no watermark, and no account at all up to 60 pages — so you can see exactly what you are dealing with before deciding anything.
Last updated . Competitor prices are as published by each vendor.