Extract Text from PDF

Extract text from PDF documents, and read scanned pages with OCR (English or Arabic) on your device. Copy it or download it as plain text.

Loading tool…

How text is stored in a PDF — and when it is not

A PDF's text layer is a series of instructions that place glyphs at coordinates using embedded fonts. There are no paragraphs, no reading order, and often no spaces as you would expect them; a word can be split across several positioning commands. Extraction means reading those instructions back, mapping glyph codes to characters through the font's encoding tables, and reassembling them into readable lines.

This is why extracted text sometimes arrives with odd line breaks or scrambled column order in multi-column layouts — the original document was designed to be looked at, not parsed. It is also why some PDFs yield nothing at all: a scanned document is a picture of text, with no text layer to extract. For those pages the tool offers OCR (optical character recognition): it draws each scanned page as an image and reads the letters with Tesseract, in English or Arabic, on your device. The first time, that downloads about 3 to 6 MB depending on the language. OCR output is a reading, not a copy, so proofread it. Afterwards you can also download a searchable copy of the PDF: the pages look exactly the same, with the recognised text laid invisibly over them, so a PDF reader can search, select and copy it.

Why we built this tool

Extraction is what you reach for when you need the words rather than the document: quoting a clause, moving a report into a document you are drafting, pulling figures into a spreadsheet, or feeding content into another tool.

The source is usually a contract, an invoice, or a report that should not be handed to a stranger's server for a task the browser can do unaided. We use PDF.js, the PDF engine built into Firefox, to parse the file locally, so your document and its extracted text both stay in the tab. Nothing is uploaded, nothing is logged, and there is no page limit — which matters when the document you are mining is a 300-page tender.

Tips and common mistakes

  • If nothing is extracted, the file is a scan — a picture of text — and needs optical character recognition, not extraction.
  • Multi-column layouts commonly extract in the wrong reading order; check those pages first.
  • Ligatures and typographic quotes can come through as unusual characters, so run a cleanup pass before reusing the text.
  • Tables lose their structure almost always; expect to rebuild them rather than paste them.
  • If a page comes back empty, it is probably a scan: choose the language and press “Read scanned pages (OCR)”.

AI on this page: Reading scanned pages uses Tesseract OCR (Apache 2.0), an AI text reader that runs on your device: about 3 MB (Arabic), 4.5 MB (English) or 6 MB (both) is downloaded only when you press the OCR button, and the PDF is not uploaded. It reads clear printed pages well; check names and numbers.

Quick start: using Extract Text from PDF

  1. Load the PDF you want to read text from.
  2. Run the extraction and wait for the pages to be parsed in your browser.
  3. Review the output — check column-heavy pages first, as those are where reading order goes astray.
  4. Copy the text you need, or download the whole extraction as a file.
  5. If nothing comes out, the PDF is a scan; the pages are images and contain no text layer.

Tidy the result with Remove Extra Spaces or Remove Formatting, and count what you extracted with Word Counter.

Two kinds of PDF, two different outcomes

This is the single most important thing to understand:

  • Digital PDFs — created from a word processor or export — contain real, selectable text. Extraction pulls it out accurately, character for character.
  • Scanned PDFs — photos of pages — contain no text at all, only images of text. Plain extraction returns nothing because there is nothing to copy; those need optical character recognition (OCR) to recognize the letters in the image.

If you can select and highlight text in your PDF reader, extraction will work. If your cursor just draws a box over the page, it is a scan.

Cleaning up extracted text

Even from a clean digital PDF, extracted text often needs light tidying:

  • Line breaks may appear mid-sentence because the PDF wrapped text to fit a column. Remove stray breaks so paragraphs flow.
  • Columns and tables can extract in an odd reading order, since PDFs store position, not meaning.
  • Headers, footers, and page numbers get pulled in with the body text and usually need deleting.

Tools like Remove Extra Spaces and Remove Formatting help clean the output quickly.

What to do with the text next

  • Quote or cite a passage without retyping it.
  • Translate the content by pasting it into a translation tool.
  • Move a table into a spreadsheet, then split it with a tool like CSV to JSON if you need structured data.
  • Repurpose an old report into new material.

When you want pictures instead

If you actually need the pages as images — for a slide or a preview — PDF to Images is the right tool. Extraction is only for getting editable words.

Related PDF Tools

Browse all PDF Tools →