How to Make a Scanned PDF Searchable and Selectable

1. 1

   Confirm it is scanned

   Try to select a line of text or search for a visible word.

2. 2

   Improve the source if you can

   Rotate sideways pages and use the sharpest copy available.

3. 3

   Run OCR

   Upload the file to an OCR tool and let it recognise the text.

4. 4

   Proofread

   Check important names, numbers and dates against the original.

A scanned PDF stores each page as an image. It looks like a document, but there are no characters behind it. Three quick tests tell you which kind you have:

• Select test: drag the cursor over a line. If individual words highlight, the file has a text layer. If a whole block or the entire page highlights at once, it is an image.
• Search test: press Ctrl+F (Cmd+F on Mac) and type a word you can see. No matches usually means no text layer.
• Zoom test: zoom in far. Real text stays crisp, while scanned text turns blocky or fuzzy.

Some PDFs mix both: a text-based page with scanned pages inserted. Test a few pages, not just the first.

OCR analyses each page image, finds lines and characters, and converts them to text. Depending on the tool, that text may be added as a hidden layer behind the picture, so the page looks the same but becomes searchable, or it may be exported to another format such as a Word document.

After OCR you can find words with Ctrl+F, copy passages into other documents, and use screen readers more effectively. It is useful for archived contracts, receipts, textbooks and old paperwork.

Be aware of how this tool works. OCR runs on our servers and is available on paid plans, so it is not a free browser-only feature. Files are handled securely and removed from our servers after processing, and files up to 50 MB are supported. As described on the tool page, the output is a Word document you can search, select and edit. If you need a PDF with a text layer, check what the tool returns on your plan before relying on it.

1. 1

   Upload the scanned PDF

   Drag and drop the file or browse from your device.

2. 2

   Start OCR

   Click once to run recognition. Larger files take longer.

3. 3

   Download the result

   Save the output and open it in Word or another editor.

4. 4

   Search and proofread

   Use Ctrl+F to test it, then read through the key parts.

• Start from a sharp, high-resolution scan. If you are scanning yourself, light the page evenly and hold the camera straight.
• Straighten and rotate pages before recognition, since sideways or skewed text confuses the engine.
• Prefer black text on a plain, light background. Coloured or patterned backgrounds lower accuracy.
• Avoid heavy compression before OCR. Compress after you have the text.
• Standard fonts work better than decorative ones, and very small print is harder than normal body text.
• Re-scan any page with shadows, creases or glare rather than hoping the engine copes.

OCR is built mostly for printed text. Handwriting is much less reliable and often needs retyping. Tables may come back with text in the right order but cells misaligned, and numbers in tables deserve extra checking. Multi-column pages, footnotes, stamps and text over images can be read in the wrong order or skipped.

Faint photocopies, fax-quality pages and unusual languages or scripts also reduce accuracy. Treat OCR output as a strong first draft, not a verified copy.

Spot-check names, amounts, dates and reference numbers against the original page, because a single misread digit can matter. Keep the original scan as the record, and use the recognised text for searching and editing. If you will convert the result to Word, see how formatting carries over and fix leftovers. For sensitive documents, remember that anyone with the output can read all the text, so store and share it as carefully as the original.

Not always. If you only need to read a document, a scan is fine as it is. OCR becomes worthwhile when you need to find words across many pages, copy quotes, feed the text into another tool, or make the content usable with a screen reader. It is also a prerequisite for several other tasks: converting a scan to Word, comparing two scanned versions, or extracting plain text all depend on there being real text to work with.

If you created the scan yourself from paper, think about capturing it well the first time, since OCR cannot recover detail that the camera never recorded. And if the document started life as a digital file, ask for that original instead: an exported PDF will always be more accurate than any recognition of a printout.

For large archives, test a few representative pages first. That tells you whether the quality of your scans is good enough before you spend time processing hundreds of pages, and it shows which kinds of pages, such as tables or stamps, need manual review.