How do I make a scanned PDF searchable with OCR?
Upload the scan to Sealique and choose English, French, German or Swahili: the recognised text is added as an invisible layer, making the PDF searchable in any browser.
A scanned contract, a photographed letter, an old report: as a picture it cannot be searched, quoted or converted. OCR reads the characters and puts real text behind the image, so Find works, PDF to Word has something to work with and screen readers can read the page.
Make a PDF searchable now Create an account Verify a signed document
Key facts
- OCR PDF on Sealique recognises the text of scanned pages in English, French, German and Swahili; it is included with an account, with no card and no daily limit.
- The recognised text is added as an invisible layer over the page image, so the scan looks the same but can be searched, selected and copied.
- Pages that already contain text are kept as they are by default, so a mixed document is processed in one run.
- Up to 80 pages per run; split a longer document first. Recognition runs on the server with Tesseract.
- A searchable PDF can be converted to Word or Excel, redacted by drawing boxes, and sealed with a PAdES digital signature.
How does the recognition work?
Each page is rendered to an image on Sealique servers and passed to Tesseract with the languages you chose. The engine returns the words with their positions, and the tool writes a PDF page that shows the original image with the words placed invisibly over it. Pages that already carry text are copied unchanged, so a document that mixes scans and digital pages is handled in one pass.
What becomes possible after OCR?
Find in any viewer, copying a clause into the minutes, PDF to Word and PDF to Excel conversions, redaction of names found by searching, and accessibility for screen readers. The searchable copy can then be sealed with a PAdES digital signature and verified online by its verification ID or QR code.
OCR PDF, step by step
Open OCR PDF
Sign in, open PDF tools and choose OCR PDF.
Choose the scan
Drop a PDF of up to 50 MB and 80 pages, or pick one from your library.
Pick the languages
Tick the languages the document uses; a bilingual document can use two at once. Keep Skip pages that already have text unless the existing text is wrong.
Recognise
Press Recognise text. Pages are rendered and read on the server; the searchable copy appears in your library next to the original.
OCR PDF is included with a Sealique account; the processing happens in your browser and on our servers, nothing is shared with third parties.
What OCR PDF includes
- English, French, German and Swahili
- Several languages in one run
- Invisible text layer; the scan looks unchanged
- Pages with text already are skipped
- Searchable copy saved in your library
- Works from a phone or a computer
Questions people ask
Does OCR change how the pages look?
No. The scan stays as it is; the recognised text is placed invisibly on top so it can be searched and selected.
How accurate is it?
Clean, straight scans at 200 to 300 dpi in a supported language read almost perfectly; faint, skewed or handwritten pages read worse.
Always proofread recognised text before it is quoted; the picture of the page remains the authoritative copy.
Which languages are supported?
English, French, German and Swahili, alone or combined in one run.
Kinyarwanda is not available in the recognition engine yet; Swahili and French data cover many of its letters, but results should be checked.
Is there a page limit?
Yes, 80 pages per run; split a longer scan with Split PDF and merge the results afterwards.
Why does the page say a server component is missing?
Recognition needs Tesseract and Poppler on the server; when one is absent the page says so and the administrator can install them.