Extract Text from PDF

Extract Text from a PDF — Clean, Copyable, Nothing Uploaded

Pull all the text out of a PDF and either copy it to your clipboard in one click or download it as a clean .txt file. The extraction runs inside your browser, so the document is never uploaded, never stored, and never seen by anyone. Free, no sign-up, no watermark, no limits.

Why PDF Text Is Awkward to Copy

A PDF is a description of a printed page. It records that a particular glyph sits at particular coordinates in a particular font — and nothing more. It does not record where a sentence begins, which fragments belong to the same paragraph, or in what order a human should read two side-by-side columns.

So when you drag-select and copy, the viewer hands you the fragments it finds, roughly in the order they were written into the file. The result is the familiar mess: a hard line break at the end of every printed line, hyphens left stranded inside words, and occasionally the right-hand column threaded into the left.

How to Extract the Text

1. Drop your PDF in — it is read locally, page by page, with progress shown.
2. Read the preview — the reconstructed text appears with counts for pages, words, paragraphs, headings and characters, so you can confirm the extraction worked before doing anything else.
3. Narrow it down (optional) — enter a page range like 12-18 to take only one chapter or section.
4. Copy or download — Copy all text puts the whole document on your clipboard; or save a .txt file. Word and Markdown are available too if you need structure.

Does Your PDF Actually Contain Text?

This is the one thing worth checking first, because it decides whether any extractor can help you:

• Try to select a sentence in a PDF viewer. No highlight means no text layer.
• Press Ctrl+F and search for a word you can see. No match means no text layer.
• Drop it in here. If the tool reports zero words, it is telling you the same thing.

A PDF exported from Word, InDesign, LaTeX, a browser or almost any application contains a proper text layer. A PDF produced by a scanner or a phone camera is a stack of photographs — the words are pixels, and reading them requires OCR (optical character recognition), which this tool does not do. Some scanners add a hidden text layer automatically; if yours did, the text will come out here.

What People Use This For

• Quoting and citing — lifting an exact passage out of a paper, judgment or report without retyping it.
• Note-taking and study — moving a chapter into Notion, Obsidian or OneNote.
• Translation — feeding clean prose to a translator instead of line-broken fragments, which translate badly.
• Accessibility — getting text into a screen reader or a plain, resizable format.
• AI assistants — pasting a document into ChatGPT, Claude or Gemini; the extraction stays on your machine, so only what you choose to paste goes anywhere.
• Reuse and republishing — moving copy out of an old PDF brochure into a CMS.

Known Rough Edges

• Hyphenated line breaks stay as they are, deliberately — guessing would corrupt genuine compounds like “well-known”. A find-and-replace fixes them in seconds.
• Multi-column magazine layouts can interleave, because the PDF stores no reading order. Single-column documents come out clean.
• Headers, footers, page numbers and footnotes appear inline, since a PDF does not label them.
• Tables arrive as lines of text, not as a grid.

Frequently Asked Questions

Why is copying text straight out of a PDF so messy?

Because a PDF does not store sentences - it stores individual glyphs at fixed coordinates. When you select and copy in a viewer, you get whatever fragments fall inside the selection rectangle, in whatever order they were written to the file. That is why pasted PDF text arrives with a line break after every visual line, hyphens stuck in the middle of words, and columns sometimes interleaved. This extractor groups the fragments back into lines, and lines back into paragraphs, before giving you the text - so it pastes as prose rather than as a column of stubs.

How do I know whether my PDF even has text to extract?

Three quick tests. First, open it in any viewer and try to drag-select a sentence: if no highlight appears, there is no text layer. Second, press Ctrl+F and search for a word you can see on screen: no match means no text. Third, drop it into this tool - if it reports zero words, the PDF is a scan. In that case the pages are images and you would need OCR, which this tool does not perform.

Does extracting text change or damage my original PDF?

No. The file is opened read-only in your browser's memory and nothing is written back. Your PDF on disk is untouched, and closing the tab discards everything.

The Copy button did not work - what happened?

Browsers only allow clipboard writes in response to a direct user action, and some block it entirely on insecure connections or with strict privacy settings. If the copy fails, the tool tells you so; you can then select the text in the preview panel and copy it manually, or download the .txt file instead.

Why do some words have hyphens in the middle?

Typeset documents hyphenate words that break across a line ending, and the hyphen is a real character in the PDF. Because there is no reliable way to tell a line-break hyphen from a genuine hyphen in a compound word, the text is left exactly as the PDF stored it rather than guessing and corrupting words like 'well-known'. A find-and-replace afterwards handles it in seconds.

Are accented letters, Greek and non-Latin scripts handled?

Yes, provided the PDF maps its glyphs to Unicode properly, which almost all modern PDFs do. Output is UTF-8. Very old PDFs with broken or custom encodings can produce garbled characters, and that is a property of the file rather than something a converter can repair.

Will footnotes, headers and page numbers appear in the text?

Yes, as ordinary text where they occur on the page, because a PDF does not label them as footnotes or running heads. If they are a nuisance, the page-range option lets you skip front matter, and they are easy to strip afterwards.

Can I extract text from just a few pages?

Yes - enter a range such as 12-18 and only those pages are read. Handy for a single chapter of a long report or a specific section of a contract.