Extract text from a PDF
Upload a PDF and get all its text, ready to copy or download as a .txt file. Everything happens in your browser, nothing is uploaded.
One PDF file
How it works
The tool reads the PDF's text layer and rebuilds it page by page. It works with «real» PDFs (created from Word, Google Docs, digital exports). If the PDF is a scan or a photo, the text is actually an image and can't be extracted this way: in that case you need OCR (optical character recognition).
What it's for
Handy to quickly copy a document's content, reuse paragraphs, search for words, or turn a PDF into a simple editable text file. If you have many PDFs and are looking for which one holds a word, there is Search a word in many PDFs.
Why copy-paste from a PDF comes out badly
A PDF does not contain paragraphs, it contains pieces of text with coordinates: the program that wrote it said «put this word here». There is no information about where a line ends and the next begins, and the reader has to infer it from the positions.
It is why pasting gives you broken lines, mixed columns and words split by an end-of-line hyphen. With two-column text the reading order can come out shuffled, because on the page the two blocks sit side by side and in the file they are just pieces with x and y values.
If the PDF is a scan, there is no text
A scanned document is a photograph of the page: there is no text underneath, which is why the reader's search finds nothing. The signal is simple: if trying to select a word selects a whole rectangle, it is an image.
In that case you need optical character recognition, which is a different job: it looks at pixels and tries to guess the letters. Many recent scans already carry a hidden text layer under the image, and then the text extracts normally.
What to do with the extracted text
The text that comes out nearly always needs stitching back together: joining the lines of one paragraph, removing hyphenation, throwing away repeated headers and page numbers. Doing it by hand over twenty pages is an afternoon's work, and nobody finishes it.
If the PDF contained tables, the extracted text loses them: the columns become a stream of words and the relationship between cells disappears. Those need a tool that looks at the coordinates rather than the text alone, otherwise the result is unreadable.
Nearby tools
To tidy up copied text there is Stitch text copied from a PDF, and for tables Extract tables from a PDF. If the file is a scan, the way through is Image to text.