PDF to Text

Pull the text out of a PDF and copy it or save it as a .txt file. Runs entirely in your browser — the file is never uploaded.

  • 100% in your browser
  • No upload, no signup
  • No page or size limit
  • Files never leave your device

If your PDF is a scan, there is no text inside it to extract. We check every page and tell you before you download anything.

Why people pull the text out of a PDF

A PDF is a description of a printed page. It knows where every character sits, what size it is and which font draws it — but it does not know that those characters form a sentence, a heading or a paragraph. That is fine when you want to read the page, and a problem the moment you want to do anything else with what it says.

So people reach for plain text. Someone with a long report wants to search it properly, or jump between its sections instead of scrolling through ninety pages. Someone is quoting a paragraph into an email and does not want the font, the line breaks and the invisible formatting to come with it. Someone is pasting a document into an AI assistant, which wants words and not page furniture. Someone is archiving a contract in a format that will still open in twenty years without any particular software.

There is also the accessibility case, and it is more common than it looks. Multi-column layouts are hard to read on a phone, and a screen reader following the wrong order reads a two-column page as nonsense. Stripping a document down to its words in the right order makes it usable in ways the original layout prevents.

Upload toolsTheir server?This toolYour deviceNothing leaves your device, so there is no copy to delete.
Other converters upload your document. This one reads it where it is.

Two kinds of PDF, and only one of them has text

Extracting text from a PDF means reading the characters stored inside the file. That works when the PDF has a text layer — the actual letters, kept alongside the instructions that draw them. A scanned PDF has no text layer: it is a picture of a page, and there are no characters in it at all.

Most PDFs made by software — exported from a word processor, printed to PDF from a browser, generated by an invoicing system — carry a text layer. You can tell by opening the file and trying to select a sentence with your cursor. If a highlight follows the words, the characters are really there and any tool can read them.

A PDF made by a scanner or a phone camera is different. Each page is one image. The words are visible to you because your eyes do the recognising, but as far as the file is concerned the page contains a photograph and nothing else. Selecting text does nothing, searching finds nothing, and extraction returns nothing — not because the tool failed, but because there is nothing there to return.

Turning a picture of words back into words is a separate job called OCR, and it is a recognition problem rather than a reading problem: software guesses at each shape and is sometimes wrong. We do not do it here. What we do instead is check every page and say which category it falls into, before you download anything. A tool that hands you an empty file, or that quietly switches to OCR without telling you, has left you unable to tell which kind of document you had.

The check is not a guess. A page with a text layer reports its characters; a page without one reports nothing at all. There is no middle ground to get wrong, which is why the result appears above the text rather than buried in a warning somewhere.

Invoice 2026Total 1,240.00Due 30 daysReal textJust pixels
Two PDFs that look identical on screen: one holds characters, the other only a picture of them — and zooming in never turns a picture back into words.

How to extract text from a PDF

Nothing is installed and nothing is uploaded. Drop the file in, look at what we found, then take the text however you need it.

Drop a PDFCheck pagesCopy or save
Drop the file, read what we found, then copy or save the text.
1

Drop your PDF in

Choose a file or drag it onto the box. It is read by your own browser — it is not sent anywhere, and there is no page or size limit beyond what your device can hold.

2

Read the summary first

Before any text appears you are told how many pages carry a text layer, which pages are scans, whether the document has filled form fields, and whether a two-column layout was found and reordered.

3

Adjust the reading if you need to

By default the text follows the page as you see it and loose lines are joined back into paragraphs. Both can be switched: document order preserves the sequence the PDF stores internally, and keeping every line as-is preserves the original line breaks.

4

Copy it or save it

Copy puts the text straight on your clipboard, which is what most people want when they are pasting into a document or an AI assistant. Download writes a UTF-8 .txt file with the same content.

Related: If you only need the text from a few pages, it is quicker to extract those pages first.

What this PDF to text converter does differently

The naive way to extract text is to take the list of pieces the PDF hands over and join them together. It is one line of code and it produces something that looks almost right — which is why so many tools do it. On a two-page tax form that approach glues words together across seventy-six line breaks. On eight pages of a research paper, two hundred and ninety-three. You get "permission toreproduce" and "neural networksin particular", scattered through the document where you will not necessarily spot them before you use the text.

The deeper problem is order. A PDF stores text in the order it is painted onto the page, which is not the order you read it in. On a form, a date field can come out as "MM DD YYYY/ /" because the separators are drawn last. On a two-column page the two columns interleave, one line each, which is exactly what happens when you select a whole page in a PDF reader and paste it somewhere. We rebuild the reading order from where the characters actually sit, which gives back "MM / DD / YYYY" and keeps each column whole.

Then there are the things that are easy to miss entirely. If you have filled in a PDF form, none of your answers are in the page text — they live on the form fields, and an extractor that only reads the page hands you a blank template while looking like it worked. We collect them and add them at the end. And some PDF generators emit characters that look right but are not: a Chinese document can come out full of Kangxi radicals, indistinguishable on screen, invisible to a search, wrong when pasted. Every string we hand over is normalised so those fold back to the characters they are drawn as.

What people use it for:

  • Searching a long report properly, or jumping between its sections
  • Pasting a document into an AI assistant without the page furniture
  • Quoting from a paper without dragging the formatting along
  • Getting the answers back out of a form somebody filled in and returned
  • Archiving a document as plain text that will open anywhere, forever
  • Reading a multi-column layout on a phone, in one readable column

Works wherever you already are

The work happens inside your browser, so there is nothing to install and no account to make. Any current browser on any operating system will do.

§ 01

Windows

Chrome, Edge or Firefox. No reader, no Acrobat subscription, nothing to download.

§ 02

macOS

Safari, Chrome or Firefox. Faster than opening the file in Preview and selecting page by page.

§ 03

Linux

Any modern browser. The same result as the pdftotext command line, without leaving the page.

§ 04

iPhone and iPad

Safari or Chrome. Useful when a PDF arrives by email and you need the text in a message.

§ 05

Android

Chrome, Firefox or Samsung Internet. Handy for reading a two-column document in one column.

§ 06

Chromebook

Works fully offline once the page has loaded, which suits a managed school device.

Why copying text out of a PDF usually comes out wrong

If you have ever selected a page in a PDF reader, pasted it, and found the sentences shuffled, this is why — and it is the same reason most conversion tools produce the same mess.

Draw orderReading order
A PDF stores text in the order it is painted. We rebuild the order you read in.

What the PDF stores

The order the page is painted in. Text goes down in whatever sequence suited whatever produced the file: a heading, then a footer, then the body, then the separators that fall between the fields. Copying follows that sequence, so on a two-column page you get one line of the left column, one line of the right, all the way down.

What we rebuild

The order you would read in. Characters are grouped into lines by where they sit, lines are grouped into columns, and the columns are read one after the other. Line breaks come back, columns stay whole, and separators return to the middle of the fields they belong to.

Where this still has limits

A picture of a page cannot be read at all, and a table stops being a table once it is plain text — rows survive, the grid does not. A line that spans both columns without any part crossing the gutter, like a row of author names, is indistinguishable from one line of each column, so a title block can come out in an odd order. Adobe's own community says the same thing about Acrobat: with sidebars and inset copy, it is "almost impossible" to know which line to follow. When our reading looks wrong, switch to document order and compare.

Text or Word?

Plain text is the right answer when you want the words and nothing else — searching, quoting, pasting into an assistant, archiving. It is the wrong answer when you need the document back as a document, with its headings, tables and images intact; that is a job for a PDF to Word conversion, and no .txt file will do it. It is worth knowing which one you came here for before you judge the result.

Your document stays on your device

The files people bring to a tool like this are rarely trivial. They are contracts, invoices, CVs, medical letters, filled-in application forms — documents whose whole content is the sensitive part. That is worth thinking about before uploading one to anywhere.

This tool does not upload anything. Your browser opens the PDF, reads it and assembles the text on your own machine. There is no request carrying your file, no queue on a server, nothing cached anywhere, and no copy for anyone to delete later. Close the tab and it is gone.

It also means the usual limits do not apply. There is no page cap, no file-size cap and no daily quota, because there is no server bill behind any of it — only what your own device can comfortably hold.

§ 01

No upload

The PDF is read where it already is. It never travels.

§ 02

No account

No sign-up, no email, no watermark on the result.

§ 03

Nothing kept

There is no copy on our side to store, log or delete.

Questions about extracting text from PDFs

Why is my extracted text empty?
Almost certainly because the PDF is a scan. A scanned page is an image, and there are no characters inside it to read — extraction returns nothing because there is nothing there. Open the file and try to select a sentence: if no highlight appears, that is what happened. We tell you which pages are affected rather than handing you a blank file.
What is a text layer?
The characters stored inside the PDF, alongside the instructions that draw them on screen. A PDF exported from software has one; a PDF made by a scanner or a camera does not, because it contains a photograph of a page instead. Everything about extracting text follows from which of the two you have.
Do you run OCR on scanned PDFs?
No. OCR recognises shapes and guesses at characters, which is a different job with different failure modes, and we would rather not pretend otherwise. If your pages are scans we say so and point you at turning them into images, which you can then run through an OCR tool of your choice. What we will not do is quietly substitute a guess for the thing you asked for.
Why does copy-paste from a PDF come out jumbled?
Because a PDF stores text in the order it is drawn, not the order you read it. Copying follows the stored order, so a two-column page comes out with the columns interleaved a line at a time, and fields drawn in an odd sequence come out scrambled. This tool rebuilds the order from the position of the characters on the page, which is why the result usually reads correctly where a plain copy does not.
Why is the order still wrong in my two-column document?
Column detection works from the gap running down the middle of the page, and some layouts do not give it a clean answer — a wide figure, a sidebar, or a row of names spread across the full width. If a page comes out in the wrong order, switch the reading order to "document order": that gives you the sequence stored in the PDF, unmodified, which is sometimes the more useful of the two.
I filled in a PDF form — why is my text missing?
Because what you typed is not part of the page. Form values live on the form fields, separately from the text that was there when the document was created, so a tool that only reads the page gives you a blank template and no warning. We read the fields too, and add the answers at the end of the text under their own heading. You can switch that off if you only want the page itself.
Will tables keep their layout?
No, and no plain-text tool can promise otherwise. A .txt file has no notion of a table — the cells come out as text in reading order, row by row, without the grid. Column boundaries usually survive as spaces, which is enough to see where one field ends and the next begins, but if you need a real table you want a spreadsheet or Word conversion instead.
Should I convert to text or to Word?
Text if you want the words: searching, quoting, pasting into an AI assistant, archiving something that will open anywhere. Word if you want the document back as a document, with headings, tables, images and styling. They are genuinely different jobs, and a .txt file that has thrown away the formatting is doing exactly what it is supposed to do.
Are my files uploaded?
No. The PDF is opened and read by your own browser, and the text is assembled on your device. Nothing is sent to a server, so there is no copy anywhere for anyone to keep or delete. That is worth having for a tool whose typical input is a contract, a CV or a filled-in form.
Is there a file size limit?
There is no limit we impose, because there is no server involved. The practical ceiling is your own device: a very large document uses memory while it is being read, and an old phone will struggle with a file a laptop handles easily. Pages are processed one at a time so progress keeps moving on big documents.
Does it work with Chinese, Japanese, Korean or Arabic PDFs?
Yes. Text in any script the PDF carries is extracted the same way, and the output is UTF-8. There is one detail worth knowing about East Asian documents: some PDF generators emit Kangxi radicals in place of ordinary characters — they look identical, but a search for the normal character will not find them. We normalise those back before you get the text, which is not something most tools do.
What encoding is the .txt file?
UTF-8, without a byte-order mark. That is what any current editor, terminal or programming language expects, and it keeps accented, Cyrillic, Greek and CJK text intact. If you are used to the pdftotext command line, this is the same output convention.

Get the text out of your PDF

Free, unlimited, and processed entirely in your browser — with an honest answer about what is actually in your file.

Extract text from a PDF