Extract and copy text from your PDF file
Pull the text out of a PDF, including scanned pages where the words exist only as pictures. Upload your document and get back text you can select, copy, and search. Free, runs in the browser, and requires no software or account.
Two PDFs can look completely identical on screen and be entirely different underneath, and understanding which kind you have explains everything about what you can do with it.
A digital PDF — one exported from Word, a design tool, or a web page — contains real text. The characters are stored as data with fonts and positions. You can select a sentence, copy it, search for a word, and the reader finds it instantly.
A scanned PDF contains a photograph of a page. To your eye it shows words; to the computer it is a grid of coloured pixels with no more textual meaning than a picture of a tree. Nothing can be selected, nothing can be searched, and nothing can be copied.
OCR bridges that gap. It examines the image, identifies the shapes as letters, and produces actual text from them.
Open the PDF and try to drag-select a line of text. If it highlights, you have a digital PDF with a text layer already. If your cursor draws a rectangle over the page but nothing highlights, it is a scan and OCR is what you need. This check takes two seconds and immediately tells you which tool to reach for.
OCR is pattern matching, so the cleaner the pattern the better the result.
Resolution matters most. At 300 DPI the letterforms are well defined and recognition is reliable; at 150 DPI characters blur together and errors multiply. If you control the scanning, use 300 DPI or higher.
Straightness matters more than people expect. Even a few degrees of skew degrades results noticeably, because the algorithm looks for text along horizontal lines.
Contrast matters. Crisp black on white recognises well; faded photocopies, coloured paper, and light grey printing all cause problems.
Typeface matters. Standard serif and sans-serif fonts are what OCR is trained on. Decorative, script, and heavily stylised fonts are much harder.
Cleanliness matters. Dust, staple marks, coffee rings, and dark scanner borders all introduce noise. Running the file through Crop PDF first to trim away black edges often improves the outcome measurably.
Students and researchers pull quotations out of scanned books and journal articles instead of retyping them. Offices digitising archives make decades of paper records searchable. Legal teams search scanned discovery documents for specific terms. Accountants extract figures from scanned invoices and statements. Anyone who has been sent a scan and needs to quote from it or work with the content.
Upload your PDF by dragging it onto the box or clicking to browse. The document is analysed and the recognised text is returned for you to copy and use. Longer documents take proportionally more time, since every page has to be processed individually.
Always read through the result before relying on it. OCR is very good but not perfect, and the errors it makes tend to be plausible-looking substitutions — a lowercase l becoming a 1, or rn being read as m — rather than obvious nonsense. A quick proofread catches these.
OCR is built for printed text, where every instance of a letter looks essentially the same. Handwriting varies between individuals, and within the same person's writing from one line to the next. Neat, well-separated block capitals occasionally produce usable results, but ordinary cursive writing generally does not. This is a limitation of the technology rather than of any particular tool.
Your document is processed in memory and discarded the moment the operation finishes. Nothing is written to disk, nothing is logged, and no copy is retained — worth knowing when the scanned material is confidential correspondence or personal records.
PDF to Word produces an editable document from a PDF that already has a text layer. Crop PDF removes scanner borders before recognition. Rotate PDF straightens pages that scanned sideways, and Translate PDF converts the content into another language.