Make scanned PDFs searchable with OCR
An image-only scan can look perfectly readable but contain no selectable words. OCR adds recognized text so searching, selection and text-based tools can work.
Start with a readable scan
Use an in-focus page with sufficient contrast. For camera captures, the photo can include the background, but all four paper corners need to be visible so you can select the complete page. Straightening cannot recover a cut-off corner.
Fictional example: photograph a two-page meeting handout. Fit the corner handles to each sheet, check the page order and zoom in to confirm the smallest printed lines are readable.
- Use Scan documents for photos, or start with an existing scanned PDF.
- Review the crop, orientation and page order.
- Open OCR PDF with the scanned document.
Select recognition settings
Choose the language that matches the source and select the pages to recognize. Check the processing badge and cloud choice: available processing modes determine whether the selected job stays on the device or uses temporary cloud storage.
Recognition quality depends on the source. Blur, skew, faint print, handwriting and complex columns can produce missed words or an unexpected reading order.
Verify that searching works
Save the searchable PDF and look for a distinctive phrase from the page. Try selecting a short line and compare it with the scan. Use Extract text to inspect the recognized wording in more detail.
The recognized text layer does not turn photographed ink into native editable page design. For a typed document, extract and correct the wording before creating a new PDF or reviewing a Word conversion.
- Run OCR and save the result.
- Search for a known heading in the exported PDF.
- Check names, dates, numbers and symbols against the page images.
When recognition is not enough
Try a clearer capture if important text is missing. Correct errors in any extracted transcript before reuse. A successful run means an output was created, not that every recognized word is accurate.
Keep the original scan as a reference. For sensitive documents, also review metadata and hidden text before sharing; OCR can create selectable text that is less obvious than the visible page image.