OCR PDF — make a scanned PDF searchable
Add a searchable text layer to a scanned PDF in 15 languages. Text becomes selectable, and tilted or sideways pages are straightened unless you switch that off. Runs on our server; files are deleted automatically within about two hours.
Drop your files here
.pdf · up to 300 MB · up to 30 jobs an hour · uploaded to our server and deleted after processing; the result is wiped within about two hours
How to OCR a scanned PDF
- 1
Upload the scanned PDF
A document that is really a picture of text, up to 300 MB and 500 pages.
- 2
Choose the document language
Pick up to three languages if the document mixes them, and decide whether to skip pages that already have text.
- 3
Download the searchable PDF
Same pages in the same order, but the text can now be selected, copied and found with Ctrl+F.
The picture stays, a text layer goes underneath
OCR does not turn your scan into a different kind of document. It recognises the words on each page and places an invisible, precisely positioned text layer behind the image you already see. Open the result and it looks like the original, except that tilted pages come back straightened (see below), and the text can now be selected, and a search engine, a screen reader or your own Ctrl+F can read it. Nothing about the page count or the order changes.
Three modes for three situations
Skip existing text is the default and the safest choice: pages that already have a text layer are left alone, and only the ones that are pure image get recognised. Redo OCR replaces a text layer that is already there, which helps when an earlier OCR pass did a poor job; it does not work together with straightening, so switch off Straighten skewed pages under More settings before you run it. Force OCR ignores any existing text and re-recognises every page as an image, which matters when the existing text layer is present but wrong, garbled or left over from an incorrect recognition pass.
Deskew, rotate and clean up before recognising
A scan photographed at a slight angle or held upside down badly hurts recognition accuracy, so deskewing and auto-rotation run by default before Tesseract ever sees the page, and the straightened page is the one that goes into the result; switch both off under More settings to keep every page exactly as it was. A background-cleanup pass is available separately for scans with visible paper texture or bleed-through from the other side, at the cost of a slightly longer run.
Processed on our server, deleted automatically
Recognising a multi-page scan reliably needs a real OCR engine running server-side rather than a browser-only approximation, so the file is uploaded for processing. The uploaded file is deleted as soon as the job finishes. The result stays downloadable for one hour and is wiped from the server within about two hours, whether or not you download it. There is no account and no manual cleanup.
Frequently asked questions
How long is my file kept on the server?
The uploaded PDF is deleted as soon as the job finishes. The searchable PDF stays downloadable for one hour and is wiped from the server within about two hours, whether or not you download it.
Which languages are supported?
Fifteen, and you can pick up to three at once for documents that mix languages in the same file.
What's the difference between skip-text and force OCR?
Skip-text leaves pages that already have text untouched and only recognises image-only pages. Force OCR re-recognises every page regardless of whether it already has text, which helps when an existing text layer is wrong or garbled.
Why does recognition come out poorly on a photo taken with a phone?
Recognition accuracy depends heavily on a flat, well-lit, evenly-angled source. Turning on deskew and rotation helps, but a photo with heavy shadow, blur or a strong angle will still recognise worse than a flatbed scan.
Does the appearance of the pages change?
Only where recognition needs it. By default a tilted scan is straightened and a page lying sideways or upside down is turned upright, and those pages come back that way. Switch off Straighten skewed pages and Auto-rotate pages under More settings to keep every page as it was; Force OCR also turns every page into an image.
Can I get plain text instead of a PDF?
Yes. Turn on the sidecar text file option and you get a ZIP with both the PDF and a plain .txt file.
Why isn't there a PDF/A option here?
PDF/A has its own dedicated tool, PDF to PDF/A, which handles the additional archival requirements (embedded fonts, colour profiles, validation) that plain OCR doesn't need.
Guides
- Compressing a Scanned Book: What Each Level Does to 60 PagesA forum question asked how to shrink a 580-page book scan without hurting the text. We ran a 60-page scan through every level of Compress PDF and checked the pixels, the sizes and the words.
- PDF to Word on a Mac Without Acrobat: Three Ways and What Each KeepsOn a Mac, Preview copies text and Pages cannot open a PDF, so the built-in route ends in copy and paste. Here is what each route keeps, from Word's own conversion to an online tool that rebuilds the file as DOCX on our server.
- Turning a Stack of Scanned Pages into a Searchable PDFA phone camera and a flatbed scanner solve two different problems, and neither one makes the result searchable by itself.
- Turning a Stack of Photos into One Properly Ordered PDFA folder of photographed receipts or pages becomes a usable document in two separate steps: assembling it, and then getting the order right.
- Building a File-Processing Chain You Can Reuse Every TimeMost real document jobs need three or four separate tools run in a specific order, and doing that from memory every time is how a step gets skipped.
- Why a Repaired PDF Can Still Fail Archival ValidationFixing a PDF that won't open and making it archive-ready are two completely different standards. Passing one says nothing about the other.
- PDF to Word and Back: What Actually Survives the Round TripConverting a PDF into Word and converting Word into a PDF are not mirror images of the same operation. One has to guess, the other doesn't.
Related tools
PDF to WordTurn a PDF back into an editable DOCX, DOC, ODT or RTF file, with OCR for scans that have no text layer at all. Runs on our server; files are deleted automatically within about two hours.
PDF to TextPull the text layer out of a PDF and save it as a plain .txt file, with the reading order and the columns intact.
Compress PDFMake a PDF smaller by re-encoding the images inside it, dropping duplicated objects and clearing metadata, without sending the document anywhere.
PDF to ExcelPull tables out of a PDF into a real XLSX workbook, one sheet per table, with an accuracy score for each. Runs on our server; files are deleted automatically within about two hours.
PDF to PDF/ATurn a PDF into PDF/A-1b, 2b or 3b for long-term archiving, then read the veraPDF validation result in plain text rather than behind a checkmark. Runs on our server; files are deleted automatically within about two hours.