EPUBLabWhich PDFs convert well, and which don't yet
EPUBLab converts PDFs with selectable text: single- and multi-column books, papers and reports, with footnotes, chapters, Arabic and Japanese. Tables are read as plain text, formulas are not rebuilt and scanned PDFs need OCR, which is not available yet.
Updated:
How this table was made
Every row comes from the documents in our public benchmark, converted with the engine build of 2026-11-03 that the site uses, and checked by hand. A row says "Not tested yet" when no document in the corpus covers it, rather than guessing.
The free sample of up to 5 pages is the real test: it uses the same engine on your own file before you pay anything.
| Kind of PDF | Status | Evidence |
|---|---|---|
| Books and long reports with running headers | Converts well | Removing repeated headers: the 9/11 Commission Report |
| Three-column documents | Converts well | Three-column PDF to EPUB: the Federal Register |
| Two-column papers | Converts, with limits: another engine keeps more continuous text on our test paper | Two-column PDF to EPUB: an academic paper |
| Footnotes | Converts, with limits: linked when the note's marker matches a call in the text | PDF footnotes to EPUB notes: an academic paper |
| Chapters and table of contents | Converts: from the PDF outline or repeated headings | Chapters and table of contents from a PDF book |
| Arabic and other right-to-left scripts | Converts, with limits: mixed-direction runs may need review | Arabic PDF to EPUB: right-to-left text |
| Japanese, horizontal text | Converts: lines joined without spaces | Japanese PDF to EPUB |
| Japanese, vertical text | Not tested yet | — |
| Tables | Not yet: read as text in reading order | — |
| Mathematical formulas | Not yet: written as plain text, not MathML | Chapters and table of contents from a PDF book |
| Fonts without a Unicode map | Flagged with a warning: affected characters show as � | — |
| Scanned pages | Not yet: flagged with a warning and left out until OCR is available | — |
| Password-protected or encrypted PDFs | Refused: remove the protection first | — |
FAQ
Is it free?
Every file gets a free sample of up to 5 pages before you pay anything, so you can judge the result on your own book. The Free plan converts up to 100 pages a month; Pro and Business convert more.
Does it work with scanned PDFs?
Not yet. EPUBLab starts with PDFs that have selectable text. Scanned books need OCR, which comes later as a separate option and will never be sold as perfectly accurate.
Will the EPUB be perfect?
No converter can promise that, and we don't. When the engine detects problems (unreadable glyphs, doubtful reading order) it shows a warning instead of hiding it, and the sample shows you the hard parts before you pay.