Extract plain text from PDFs, copying digital pages exactly and running OCR only on scanned pages, with a page range for long files. Nothing is uploaded, and there’s no account or watermark.
Drop one or more PDF files onto the drop zone, or tap “or tap to choose files”.
Pick the “Text language” of the document. In “PDF pages”, leave the box empty for the whole file or type a range such as 1-3, 5, 8-.
Press “Extract text”. Pages that already contain text are copied instantly; scanned pages are recognized one by one, with the page number and percentage shown.
Open “Show text” to see the result with a “— Page 3 —” heading before each page, then “Copy text” or “Download .txt”.
Some PDFs are born digital: exported from Word, generated by a website, or produced by accounting software. Their pages contain a real text layer, the actual characters with their fonts. Others are only pictures of pages: a scanner or phone app saved each page as an image and wrapped them in a PDF. Selecting text works in the first kind and does nothing in the second.
The tool checks every page separately. A page with a text layer is copied directly from the file, which is exact and instant, with no recognition and no guessing. A page without one is rendered at about 230 dpi and passed to the Tesseract OCR engine. A mixed file, such as a typed report with a scanned signature page appended, is handled correctly without you doing anything, and the result says how many pages came from each source.
To keep the browser responsive, up to 50 scanned pages are recognized per PDF in one run (pages with a text layer: up to 1000). If a document is longer, you get a notice, and you can continue by typing the next range, for example 51-100, in “PDF pages”.
Mac
Windows
This sends the file to Google's servers and needs a Google account. Keep confidential PDFs local.
Android
Reading the text layer is always preferable when it exists: it reproduces the document exactly, including every number, in a fraction of a second. OCR is the fallback for pages that have no text, and on a clean scan it is very accurate, but it remains an estimate that deserves a quick check on figures and names.
Sometimes copying from a PDF gives gibberish, strange symbols or letters in the wrong order, even though the text looks fine on screen. This happens when the fonts inside the file don't map their shapes to real characters. Tick “Recognize pages that already contain text” to ignore that broken layer and read the page visually instead.
The output is plain text: no fonts, columns, images or table cells. That is ideal for searching, quoting, feeding into another program or translating, but not for editing a document while keeping its look. Forms come out as their labels and filled-in values in reading order, and tables as lines of text with spaces between values.
Try to select a word in your PDF reader. If you can highlight individual words, the page has a text layer. If the whole page highlights as one picture, or nothing happens, it is a scan. You don't need to know in advance: the tool detects it per page.
Pages with real text, yes, up to 1000 at once. For scanned pages, 50 are recognized per run; type 51-100 in “PDF pages” and run again, and so on. Each run takes a few seconds per page.
The PDF's embedded fonts lack a proper character map, so the stored text doesn't match what you see. Tick “Recognize pages that already contain text” and extract again to read the pages with OCR.
No. Open the file in a PDF reader with the password, save or print an unprotected copy, and drop that copy instead.
No, the result is a .txt file. Paste it into Word or Google Docs if you want to rebuild the formatting. Choosing “Join into paragraphs” removes the hard line breaks that come from the page width.
Yes. Open this page in your phone's browser, tap “or tap to choose files” and pick the PDF from your files. Recognition runs on the phone, so long scanned documents take longer than on a computer.
Last updated October 28, 2026
Enlarges small text, turns dark backgrounds light and evens out shadows. Try it off if letters come out broken.
Up to 50 scanned pages are recognized per PDF.
Pages with real text are copied exactly and instantly, without OCR. Tick this if that text comes out garbled.