Not All PDFs Contain Text
A PDF exported from Word contains text. A PDF made by a scanner contains pictures of text. To a basic reading app, the second one is silent — there's nothing to read. Old course readers, archived reports, library scans, photographed contracts: a huge share of the PDFs people actually need to get through fall in this category.
OCR, Locally
VocaRead recognizes scanned pages with Apple Vision OCR, on your device. The pages never leave your iPhone: no upload, no cloud service, no account. That matters for a work document, and it means OCR works in airplane mode too. Recognition runs during import, page by page, and the result is a normal book in your library, with highlighting following the voice.
What Got Better in 2.6 and 3.0
Reading a PDF aloud well is mostly about not reading the wrong things. Recent versions made the extraction a lot smarter:
- Complex layouts. Two-column papers are read column by column, not line by line across the gutter.
- Headers, footers, page numbers are detected and skipped instead of being read out on every page.
- Chapter detection finds headings in the PDF and turns them into a chapter list, so you can navigate by chapter like in an EPUB — without losing prose that merely looks like a heading.
- Hyphenation and ligatures are repaired before speech: "under-standing" becomes "understanding", "fi" becomes "fi".
- Introductions and prologues are now detected, so playback starts at the actual beginning.
How to Import a Scan
- Tap + in the library and choose the PDF from Files, or use Share › VocaRead from Mail, Safari, or a scanning app.
- VocaRead detects that the pages are images and runs OCR. A progress indicator shows the pages being recognized; a 300-page scan takes a few minutes on a recent iPhone.
- Pick a voice and press play. Everything else — speed, sleep timer, bookmarks — works as with any book.
Getting Better Results
- Scan straight and well-lit. The iPhone's built-in document scanner (Notes or Files) produces excellent input.
- Prefer 300 dpi when you can choose. Tiny text in low-resolution scans is where OCR errors cluster.
- Fix names once. If a recurring name is misread or mispronounced, add a pronunciation rule and it applies to the whole book.
- Combine with highlights. Mark passages while you listen and export them as Markdown for your notes.
For regular, text-based PDFs, see how to turn any PDF into an audiobook. Ready to try it on that stack of scans? Download VocaRead — 90 minutes of free listening per document.