OCR a scanned PDF for free and make it searchable
A scanned PDF is a stack of pictures of pages. You can read it, but you cannot select, copy or search the words, and copying gives you nothing. OCR (optical character recognition) looks at each page image and works out the letters on it.
Most online OCR services upload your document to their servers, and free plans often cap you at a few pages or a few files a day. The tool on this page runs the open-source Tesseract engine inside your browser instead, so the scan of a contract, a medical letter or an ID never leaves your device, and there is no daily limit. You can get a searchable PDF that looks exactly like the original with an invisible text layer on top, the plain text, or both.
Pages are drawn with pdf.js and read with Tesseract, on your device. English is built in.
Runs in your browser. Your file is never uploaded.How we keep your files private
How to OCR a scanned PDF and make it searchable in your browser
- Check the PDF really is a scan: if you can already select the text, use PDF to text instead, which is faster and exact.
- Drop the scanned PDF on the box above (up to 50 pages and 50 MB).
- Choose the output: Searchable PDF + text (the default), Searchable PDF, or Text (.txt). Leave Text layout on Auto for most letters and documents.
- Click Read text (OCR) and wait. Each page takes a few seconds; the first run also loads the reading engine (about 7 MB), which is cached afterwards.
- Download the searchable PDF and check it with Ctrl+F (Command-F on a Mac); copy or download the text if you need it, checking names and numbers against the scan.
How to get more accurate OCR
- Scan at 300 DPI if you can. Below 200 DPI, small print starts to break up into wrong letters.
- Turn sideways pages upright first with Rotate PDF. OCR reads upright text best.
- Use black-and-white or greyscale for plain text documents, and avoid shadows and curved pages when you photograph a book.
- Split very long documents with Split PDF and read them in parts of up to 50 pages.
What OCR cannot do well
- Handwriting
- Tesseract is trained on printed text. Neat capitals sometimes work; joined-up writing usually does not.
- Other languages
- The reader built into this tool uses English language data. Text in other alphabets, and many accented letters, will not be read correctly.
- Complex layouts
- Tables, forms and multi-column magazines come out as lines of text, not as a spreadsheet or a laid-out page.
Free alternatives already on your computer
On a Mac, Live Text in Preview lets you select text in a scanned page one page at a time. Google Drive can open a scanned PDF as a Google Doc with the text recognised, but that means uploading the file to Google. Both are fine for documents that are not sensitive; the browser tool above is the option that keeps the file on your device.
FAQ
- Is this OCR really free?
- Yes. The engine runs on your own device, so there is no server cost to pass on: no sign-up, no page credits and no daily limit. The caps of 50 pages and 50 MB per file only keep the browser tab from running out of memory.
- Is my scanned PDF uploaded?
- No. Pages are drawn and read inside your browser. After the engine has loaded once, the tool works even with Wi-Fi turned off.
- How accurate is it?
- Clean printed scans at 200–300 DPI usually come out very well. The result shows a confidence score; below about 60%, expect mistakes and read the text carefully.
- Will the searchable PDF look different from my scan?
- No. The original page images are kept and the recognised words are added as an invisible layer, so the file looks the same but you can search, select and copy text. Pages that already had real text are left as they are.
- My PDF already has selectable text. Do I need OCR?
- No. Use PDF to text, which copies the existing text layer exactly and much faster.
Related
- PDF OCR — the full tool page, with format notes and troubleshooting
- Copy text from a PDF on a Mac
- How to redact a PDF properly
- Compress a PDF without losing quality