Turn any PDF into plain text

Two kinds of PDF, two different problems

A PDF can hold its text in one of two ways, and which one you have decides almost everything about how hard it is to get the text out.

A PDF saved from Word, Google Docs or a design program carries the actual letters inside it. Getting that text out is instant, because it's already there.

A scan or a photo holds only a picture of the text. The letters have to be recognised in that picture before they can be pulled out, and that step is called OCR (optical character recognition).

Scans take a little longer, and the result is very accurate but not perfect. We'd rather say that plainly than quote a number we can't stand behind.

Copy text from a scanned PDF

If your PDF viewer can't select or copy any text, then it's a scan. Upload it here and this online OCR reader will give the text back to you as something you can copy, search and edit.

Setting the language is often helpful. The auto-detect option works well on most pages, but choosing English, Urdu, Arabic, Chinese or any other one of 39 languages yourself yields noticeably better results on mixed or non-Latin documents.

• Scan at 300 DPI or higher. Below that, fine detail gets lost and OCR makes more mistakes.
• Straighten crooked pages before you upload them. Even a few degrees of tilt makes the result worse.
• Flat scans produce better results than photos taken at an angle.
• Good, even lighting is more important than you think: a shadow across the page makes letters harder to read.

Finally, do you want the PDF itself to become searchable, looking exactly the same as it does now? That's a different operation: PDF OCR adds a hidden text layer to the PDF, rather than outputting a separate text file.

When the text comes out jumbled

Academic papers, magazines and newspapers are set in columns. A converter that reads straight across the page mixes the columns together, and the result makes no sense.

This PDF to text converter recognises two column pages and reads each column top to bottom before going to the next one.

• Keep line breaks ends lines where the page ends them. Columns are still read one at a time.
• Merge into paragraphs joins lines which were only broken by the width of the page. Best when pasting into a document or an email.
• Preserve layout keeps words in place on the page, so that columns and tables line up in a monospaced font.
• Single block of text puts everything on one line, to paste into a form or a spreadsheet cell.

Some PDFs are made with fonts whose letters aren't real characters, so copying from them gives nonsense. Pages like that are recognised and read using OCR instead. You can also force this for the whole document with Re-read every page with OCR.

TXT, Word, Markdown, web page or JSON

• A plain text (.txt) file is the words, and only the words. This is the standard PDF to TXT conversion that allows pasting elsewhere or passing to another program.
• A Word (.docx) file opens in Word or Google Docs, with tables preserved as tables and new pages inserted as in the original PDF.
• A Markdown (.md) file preserves the headings and tables using Markdown syntax, for easy use in notes apps, wikis, or by developers.
• A web page (.html) file contains structured tags that you can view in any browser.
• A JSON file contains a representation of every page, with lines, positions, OCR confidence, tables, and links.

Need the exact page design as well as the words? Convert PDF to Word to preserve the page layout. Need the tables converted to spreadsheet form? Try our PDF to Excel conversion.

Tables, headers, and page numbers

Tables are converted to rows and columns rather than paragraphs of text. Tables are tab-delimited in the plain text version and appear as tables in the Word, Markdown, and HTML versions.

Remove headers and page numbers eliminates the running header and page number that appear at the top or bottom of every page in the PDF. Instead, the text flows as a single document without being divided into pages every few paragraphs.

Fix words split across lines rejoins words split at the end of a line with a hyphen. Instead of "docu-ment" you'll see "document". Mark where each page starts puts a short page marker at the beginning of every page, so you can easily follow along in the original PDF.

Locked and confidential PDFs

For a password-protected PDF, type its password in the Document password box. It is used to open the file and never stored.

For a PDF that should not leave your device at all, open Extract on this device. Your browser reads the text itself, nothing is uploaded, and no credits are used.

That mode reads the text layer only, so it suits contracts, statements and letters made on a computer. A scanned page has no text layer to read, so scans need the normal converter, which runs OCR. On-device mode can also be slower on long documents, because it uses your own computer rather than ours.

Read the result before you download it

The text opens in an editor on the page. Jump to any page, fix a word, join lines into paragraphs, change the case, copy it, or listen to it read aloud.

Words the OCR was unsure about are highlighted, so you know what to check. Links in the PDF are listed with the result. Not happy with it? Read the same file again with another layout, language or format, free, up to 5 times while it is still on our servers.

Translate the extracted text

You can machine-translate the result into English, Urdu, Arabic, Spanish, French or seven other languages as it is extracted. It is handy for understanding a foreign-language document quickly.

It is machine translation, so check anything important. Up to 120,000 characters are translated at a time, so for a long report choose a page range and translate it in parts.

Yes. Your files are deleted from our servers one hour after processing, automatically.

Nothing is stored past that hour. Nothing is read by a person. Nothing is used to train anything. Every upload and download runs over HTTPS, so your file is encrypted both ways.

If even that is more than you want, use Extract on this device. The PDF stays in your browser and never reaches our servers at all.

Your recent results are kept in your own browser, never on our servers, and you can clear them whenever you like.