How to make a scanned PDF searchable and editable (OCR)
A scanned PDF is a picture of a page, so there is no text to search or edit until you run OCR. Overtype uses an English model on your own computer to read the words off the scan and add a searchable text layer you can then edit.
If you press Ctrl+F in a PDF and nothing is ever found, the page is almost certainly a scan. The document contains an image; the words exist only as pixels arranged in the shape of letters.
OCR — optical character recognition — looks at those pixels and works out which characters they represent. Overtype runs an English model on your computer, adds a searchable text layer behind the page image, and then lets you select and edit the result.
Step by step in Overtype
1. Open the scan
Open the scanned PDF in Overtype. Old contracts, receipts and letters from a document feeder all qualify.
2. Run OCR
Choose OCR for scanned pages. Recognition runs on your computer using an English model, and produces a searchable text layer under the page image. Expect it to take longer than a normal edit — recognition is real work.
3. Check the result
Search for a word you can see on the page to confirm recognition worked. Faint, skewed or handwritten pages recognise less reliably than clean printed ones.
4. Search and edit the new text layer
The OCR result is a searchable text layer behind the page image. Select it, search it, correct mistakes, and edit paragraphs like any other PDF.
5. Export
Save the file to your computer, now searchable rather than a flat image.
Why this works on your device
Most OCR services upload your pages to a server farm to do the recognition. That is precisely the wrong thing to do with scanned contracts, medical letters or ID documents. Overtype runs an English recognition model locally in your browser, so the scan never leaves the machine it was scanned onto.
Questions
- How accurate is OCR?
- Accuracy depends on scan quality. Clean, straight, printed pages in English recognise well. Faint photocopies, skewed pages, handwriting and non-English text are harder, so proofread anything important.
- Why is OCR slower than other tools?
- Recognition analyses every page image on your own computer instead of a remote server, so it depends on your machine rather than a data centre.
- Can I edit the text after OCR?
- Yes. Once the words are recognised you can edit paragraphs, redact, or add text as you would in any other PDF.