Extract text
Why is text missing from some pages of my PDF?
By DeckleUpdated
When text extraction skips only some pages, those pages are almost always images: scanned sheets, photographed pages or pasted screenshots inside an otherwise typed PDF. PDF to text reads the text the PDF stores, so a page that is a picture comes out blank while its neighbours extract normally. Find the blank pages, run OCR on just those, and extract again.
Pages become plain text
Selected PDF pages
Order chosen in the review
UTF-8 text
Words without page formatting
Page boundary: U+000C or a blank line
How to find and fix the missing pages
- Open PDF to text and choose the PDF, or drop it onto the page.
- Keep All pages and choose Mark where each page ended under Between pages, so every page has its own block in the file.
- Choose Extract text and open the TXT file in a plain text editor.
- Look for empty blocks between page-break markers. Count markers to work out which page numbers they are: the block after the fourth marker is page 5.
- Open those pages in a viewer and try to select a line. If nothing highlights, the page is an image.
- Run the image pages through an OCR tool, then extract the recognised PDF again.
The page markers are what make this quick. With One continuous document you only see a missing stretch of text, not which page it belongs to.
Why one PDF can have both kinds of page
PDFs are often assembled from different sources, and each page keeps the kind of content it started with.
| How the PDF was made | Pages likely to be images |
|---|---|
| Typed contract with a signed page scanned back in | The signature page |
| Report with scanned appendices or receipts | The appendix pages |
| Several files joined with Merge PDFs | Every page from a scanned source file |
| Slide deck with screenshots | Slides that are one large picture |
| Form filled in, printed and scanned | Every page, though the blank form had text |
| Document with a designed cover | Often the cover only |
A page can also be partly text. A typed letter with a pasted image of a table extracts the letter and skips the table, because the table is pixels.
Fix the image pages without touching the rest
There is no need to OCR the whole document if only a few pages are scans. Typed pages already hold exact text; OCR on them gains nothing and can only add misreadings.
Run OCR on the scanned pages only. If your OCR tool lets you choose pages, pick just the image pages. Then extract the whole PDF again.
Or split, recognise and rejoin. Use Split PDF to separate the scanned pages, OCR that file, rejoin the parts with Merge PDFs, then move any page back to its original position with Organize PDF pages.
Or extract the typed pages now. If you only need the typed text, choose Choose pages and type the working pages, such as 1-4, 6-12. The scanned pages can be transcribed separately.
Check the pages that did extract
Pages with text may still need a look. A PDF that was run through OCR before you received it has selectable text with recognition mistakes in it. Typical signs are l for 1, rn for m and broken words at line ends. The extraction copies those mistakes faithfully. Native PDF text vs OCR text explains how to tell an OCR layer from typed text.
Tips
- Always extract with page markers when diagnosing. They turn "something is missing" into "pages 5 and 11 are missing".
- Search the PDF for a word on a blank page. If the viewer cannot find it, the page has no text layer.
- Zoom right in. Letters on an image page turn blocky or blurred; real text stays sharp at any zoom.
- Keep the page order the same. If you split and rejoin, check the final page count matches the original before extracting again.
Frequently asked questions
Why does PDF to text skip some pages?
Those pages have no selectable text, usually because they are scans or images. The other pages store real characters, so they extract normally.
Can Deckle extract text from scanned pages?
PDF to text reads selectable text only. A scanned page needs OCR first; after that, PDF to text exports the recognised words.
Do blank pages break the page count in the TXT file?
No. With Mark where each page ended, a page without text still gets its own empty block between markers, so the count stays correct.
Should I OCR the whole PDF to be safe?
No. Pages that already have text gain nothing from OCR and can pick up misreadings. Recognise only the pages that are images.
Related tools
- PDF to text to export each page's text with a page marker
- Split PDF to separate the scanned pages
- Organize PDF pages to put pages back in order
- Merge PDFs to rejoin recognised and typed pages