PDF to Text — Free Online Text Extractor

A digital PDF was produced by software, so it carries the characters themselves. A scanned PDF is a stack of page images produced by a scanner or a camera, so it carries no characters at all. That difference decides everything about the result. The digital file gives you exact text. The scanned file has to be recognized, and recognition is very good on clean scans and less good on crooked, blurry or badly lit ones. You can usually tell which kind you have by opening the file in a PDF reader and trying to select a line of text with your mouse. If the selection highlights words, there is a text layer.

For pages with a real text layer, yes. The characters are read out of the file rather than recognized from a picture, so no OCR runs on them and there is nothing to misread. Reading order is the one thing that can still surprise you, because the text layer stores where each run of characters sits rather than what it means, so a page with columns or sidebars may come out in an order that does not match the way your eye moves. The words themselves are the words in the file. Pages with no text layer go through OCR instead, and those carry the usual recognition accuracy.

Up to 25 pages per PDF. A longer file is still accepted: the first 25 pages are extracted and a notice on the result tells you how many pages were left out. The cap is there because the work runs on your own hardware, and a long scanned document means one OCR pass per page inside your browser tab. If you need the rest, split the file in a PDF reader and run the remaining pages as a second file.

No. An encrypted PDF cannot be opened without its password, and this page does not ask you for one or try to get around the protection. You get a clear error instead of a blank result. Open the file in the reader you normally use, authenticate the way you normally do, save a copy with the password removed and run that copy through here.

No. The file is read from your disk into the page and opened by a PDF reader that runs in your browser. There is no upload endpoint in this site to receive it. That is the whole point for the documents people actually bring to a tool like this: bank statements, payslips, tax forms, signed contracts and scans of an ID. Watch the network tab during a run and you will see the reader and the OCR model arriving, and nothing leaving with your document.

The OCR side reads English, Spanish, Romanian, Arabic, Greek, Turkish, Indonesian, Russian and Japanese. Pick the language before you extract, because the engine matches shapes against the model you choose and the wrong model reads badly. Pages that already have a text layer ignore the setting completely, since those characters are read out of the file as they are, whatever script they are written in.

Yes. Select a batch and the files are worked through in sequence rather than in parallel. Each result appears as soon as its file is finished, so you can start reading while the rest are still running. The dropzone shows the limit for a single run. Sequential processing keeps memory use predictable, which matters most on a phone.

No. The output is plain text, so a table comes back as its cell contents rather than a grid, and headers, footnotes and captions are not marked as anything. Reading order can also differ from the visual layout, because position is what the file stores and meaning is not. Straight prose comes through in the right order nearly always. A page with sidebars, several columns or a filled form is where you should read the output before you rely on it.

No, and we would rather say so here than let you find out with a file that matters. The OCR engine is trained on printed characters. Handwriting, signatures and handwritten form fields sit outside what it can do, and what comes back from them is noise shaped like text. A scanned page that mixes print with handwriting gives you the printed part and nonsense where the writing is.

Plain text. Copy it to your clipboard with one button or download it as a .txt file, with the pages separated and labeled so you can find your place in a long document. There is no word processor file, no searchable PDF and no formatting carried over. If you need the words somewhere else, paste them into a document, a spreadsheet or a translator.

That is the other direction, and it has its own page on this site. The JPG to PDF converter takes photos or scans and writes them into a single PDF, in your browser, with nothing uploaded. Use it when you have pictures and need a document. Use this page when you have a document and need the words out of it.

Yes, free with no account and no usage counter. There is no sign-up, no watermark on the text, no credits and no paid tier holding back a better version. The extraction runs on your own hardware, so there is no per-file server cost for us to pass on to you.