OCR PDF
Make scanned PDF searchable by adding a text layer you can search and copy.
Deleted from our servers as soon as processing endsHow to OCR a PDF
- 1 Add your PDFs Choose one or more scanned PDFs, or drag them into the box.
- 2 Set the language French, English, or both if the document mixes them.
- 3 Download Get your PDF with searchable text. Multiple files come in a ZIP.
Why OCR a PDF
A scanned document is just a series of images: you can't search it for a word, copy a sentence or have a screen reader read it. OCR PDF uses optical character recognition to add an invisible text layer under the image and make scanned PDF searchable: the page looks exactly the same, but its content becomes selectable.
Choosing the right language clearly improves the result, especially for accents. Recognition still depends on scan quality: a blurry, tilted or handwritten page will produce errors. Pages that already contain text are left as they are. To edit the text afterward, run the resulting file through PDF to text.
Measured results
Measured in November 2026, on a 3-page color scan at 300 dpi: pages of French text printed and then scanned, with grayish paper, grain and one slightly crooked page. Language chosen: "French".
| Measurement | Before | After |
|---|---|---|
| File size | 2.25 MB | 1.83 MB |
| Selectable characters | 0 | 5,095 |
| Words identical to the original text | 0 | 977 out of 979 |
On this scan, only two words differ from the original text. The resulting file is even a little lighter, but don't count on OCR to save space: to reduce the size, use Compress PDF. On this same scan, the "Strong" level gives 190 KB and the "Recommended" level 524 KB.
Real-life examples
- Finding information in scanned archives: once recognized, your statements, contracts or letters can be searched with your PDF reader's search, like a document created on a computer.
- Reusing a passage in an email or a report: select the text and copy it. Always check the numbers and names, where a recognition error easily goes unnoticed.
- Document in English: choose "English" under "Document language". For a document that mixes French and English, choose "French and English".
- Getting all the text at once: run the recognized PDF through PDF to text afterward.
Troubleshooting
- A page was not recognized: a page that already contains text, even very little, is left as it is. This happens with a scan where software added a page number, a stamp or a header as text. To force recognition, first turn the pages into images with PDF to JPG, combine them on A4 pages with JPG to PDF (20 images max), then run OCR again on this new file. It will only contain images: links and original text are lost.
- The recognized text has many errors, or wrong accents: check the language you chose, which matters most for accents. An English document should be run with "English", a French one with "French" or "French and English". If the scan is blurry or too small, scan it again if you can: our sample, scanned at 300 dpi, is recognized almost without a mistake.
- "This PDF is password protected": OCR cannot read an encrypted file. Remove the protection first with Unlock PDF.
- The file is rejected: each PDF must be 100 MB max, with 20 files at most per upload.
Good to know
OCR does not turn your scan into an editable document. The page stays an image, and the recognized text is placed underneath it, invisible, in the same positions as the words. That is what lets you select a line exactly where you see it. To get that text and rework it, run the recognized PDF through PDF to text. PDF to Word does not pick up this invisible layer: on our sample scan after OCR, the Word document only contained the images of the pages (checked in November 2026).
Privacy
Unlike most tools on the site, OCR does not run in your browser: your files are sent to our server and processed during the request. The temporary folder is deleted as soon as the recognized PDF is sent to you, even if an error occurs, and no copy is kept.
Frequently asked questions
Are my files safe?
They are processed on our servers and deleted as soon as the recognized PDF is sent back to you. Nothing is kept, viewed or shared.
Will my document look different?
No. The image of each page stays the same, the recognized text is added underneath, invisible.
Is PDF text recognition accurate?
Not always. It works well on sharp printed text. Handwriting and blurry or very small scans produce errors.
Which languages are recognized?
French and English, separately or together for a document that mixes both.
What if my PDF already contains text?
Pages that already have text are not processed again. Only pages made of images go through recognition.
Can I process several PDFs at once?
Yes, up to 20 files, 100 MB max each. A long document may take a little patience.
LordOfThePDF stays free thanks to you
Ads and your support fund the site. Support from €1 helps us keep every tool free.