English
How to extract text from a PDF
To extract text from a PDF, drop the file on the box above. The text appears on the right as soon as the pages are read, and you can copy it or download it as a .txt or Markdown file. Everything happens in your browser, so a contract, a bank statement or an unpublished manuscript is never sent to a server.
- Add one PDF or several. Use the sample button if you only want to see what the tool does.
- Pick plain text or Markdown, and decide whether lines should be joined into paragraphs.
- Type a page range if you only need part of the file, for example 1-3, 5, 8-.
- Check the result in the whole-document view or page by page, then copy it or download it. With several files, Download all (.zip) gives one text file per PDF.
The word and character counts above the text are for the pages you selected, which is handy for a quick count of a long report without opening a word processor.
Why copied PDF text looks broken
A PDF does not contain paragraphs. It contains pieces of text with coordinates, drawn in whatever order the program that made the file chose. Select and copy in a PDF reader and you get a line break at the end of every printed line, words cut in half by hyphens, and in a two-column article the first line of the left column followed by the first line of the right one.
This tool rebuilds the structure from the coordinates:
| Problem when you copy | What the tool does | Option |
|---|---|---|
| A line break after every printed line | Joins the lines of a paragraph, keeps the gaps between paragraphs | Line breaks |
| inter- national split over two lines | Removes the hyphen and joins the word | Fix words split with a hyphen |
| Columns mixed together | Reads the left column to the bottom, then the next one | Read columns one after the other |
| Page number and title on every page | Drops lines that repeat in the top or bottom margin | Remove repeated headers and footers |
| A table collapsed into one stream of words | Keeps rows and columns, tab-separated or as a Markdown table | Detect tables |
One caution about the hyphen fix: it joins a lowercase continuation to the word before it, so a real compound that happens to break at a hyphen, such as "well-
known", becomes "wellknown". Untick the option for documents full of compounds.
Plain text or Markdown?
Plain text is the safest choice for pasting into an email, a note app or a translation tool. Markdown is better when the text is going into documentation, a wiki or a notes app that understands it. The tool guesses headings from font size: the largest heading size becomes #, the next ##, then ###. Bulleted and numbered lists keep their markers and tables become Markdown tables. A PDF that uses one font size for everything simply has no headings to find.
Scanned PDFs and OCR
A scan is a photo of a page, so there is no text to extract. When a page has no text layer, an OCR box appears where you choose the language and read the pages. The OCR engine (Tesseract) runs in your browser. The engine and the language data are downloaded the first time you use it (a few megabytes, more for some languages), and your pages are still not uploaded.
- A clean scan at 300 DPI gives good results. A phone photo taken at an angle, with shadows, gives noticeably worse ones.
- Choose the right language. English OCR on a Spanish page loses every accent.
- Tables and multi-column layouts are less reliable with OCR than with real text. Proofread numbers in particular.
What it cannot do
Text that is part of a picture, a chart or a drawing is not text in the PDF and is skipped unless you use OCR on the page. Complex layouts with floating boxes, sidebars or text wrapped around images may read in the wrong order, and the page-by-page view is the quickest way to spot them. PDFs that were made with a damaged font encoding can give garbled letters, which you will see in the result before you download anything. Text drawn sideways is left out. Files over 150 MB are refused so the tab does not run out of memory.
Frequently asked questions
How do I extract text from a PDF without uploading it?
Drop the PDF on this page and the text appears on the right a moment later, ready to copy or download as .txt. Your browser reads the file itself, so nothing leaves your phone or computer. A 20-page text PDF takes a few seconds.
Why does text copied from a PDF have broken lines and strange spacing?
A PDF stores each line as separate pieces placed on the page, not as paragraphs. This tool puts the pieces back in reading order, joins the lines of each paragraph and fixes words that were split with a hyphen at the end of a line, so the result pastes cleanly into a document. If you need the lines exactly as printed, switch Line breaks to Keep line breaks.
Can it read PDFs with two or three columns in the right order?
Yes, in most cases. The tool looks for a clear vertical gap and reads each column from top to bottom before moving to the next one, while a full-width title above the columns stays on top. Magazine pages with text wrapped around pictures, sidebars or irregular boxes can still come out in the wrong order, so check those pages in the page-by-page view.
How do I extract text from a scanned PDF?
Use the OCR box that appears when a page has no text layer. Pick the language of the document and press Read pages with OCR. The engine runs in your browser and the first run downloads the language data once. A clean 300 DPI scan reads well, a skewed phone photo much worse, so proofread the result.
Can I get the text as Markdown with headings?
Yes. Choose Markdown (.md) and larger text is turned into headings: the biggest size becomes #, the next ## and then ###. Bullet and numbered lists keep their markers and tables become Markdown tables. Headings are guessed from font size, so a PDF that uses one size everywhere will have none.
How do I remove headers, footers and page numbers from the extracted text?
Leave Remove repeated headers, footers and page numbers ticked. Lines in the top or bottom margin that repeat on at least half of the pages, and lone page numbers, are dropped. A title that only appears on one page is kept. Untick the option if a footer you need is being removed.
Can I extract text from only some pages or from many PDFs at once?
Type pages such as 1-3, 5, 8- in the Pages box and only those are read. You can also add several PDFs and download one text file each in a single .zip. The same options and page range apply to every file.
Is there a limit on the size of the PDF?
PDFs up to 150 MB can be opened and up to 3,000 pages are read at once. Pages are processed one by one with a progress bar and a Cancel button, so a 200-page file does not freeze the tab. Password-protected PDFs ask for the password and use it only to open the file on your device.