How to Edit Text in a Scanned PDF: 3 Methods That Work

A scanned PDF is a photograph of paper, so there is no text to edit until you either cover the old words and type new ones on top, or run OCR (optical character recognition) to turn the picture into real text. Use an overlay for a quick fix such as a date, and OCR plus Word for anything longer.

Most people discover this the hard way. They open a scanned contract, click on the wrong date, and nothing happens: no cursor, no selection, nothing to delete. The PDF looks like text but behaves like a picture. This guide explains why, then gives you three ways to change the words, from a two-minute patch to a fully editable document.

Key takeaways

• If you cannot select words in a PDF, it is an image of a page and needs OCR before its text can be edited as text.
• For one or two small changes, cover the old text with a white box and type the new text over it in a PDF Editor. The page keeps its scanned look.
• For real rewriting, run OCR PDF, convert the result with PDF to Word, edit, then export back with Word to PDF.
• Recognition errors are normal. Proofread every name, number and date after OCR.
• A white box only hides text from view. To remove confidential words, use proper redaction instead.

Why you cannot select the words in a scan

When a scanner or phone saves a page as PDF, it stores one image per page. The letters on that image are just dark pixels. There is no record that "15 March" is a date, or even that it is made of characters. A PDF reader sees one picture, which is why search finds nothing and the cursor will not land.

OCR fixes that. The Society of American Archivists defines it as "the process of transforming images of characters in a document to the equivalent ASCII code for those characters" (SAA Dictionary). Once the characters exist as text, they can be searched, copied and edited. The catch is accuracy. As the US National Archives puts it, "OCR text is not always accurate, but it increases access and supports indexing and other search functionality" (NARA).

A quick test tells you what you have: open the PDF and try to highlight one word. If it highlights, the file has a text layer and a normal PDF Editor can change it directly. If not, read on.

Pick the right method for the job

Method 1: cover and retype small changes

This is the fastest route when a scan is otherwise correct. Siliph's PDF Editor cannot select the words on a scanned page, but it can place things on top of it.

1. Open PDF Editor and upload the scanned PDF (up to 200 MB).
2. Use Whiteout to draw a solid box over the words you want to replace. Size it to the line, not the whole paragraph.
3. Choose Add text and click inside the white box. Type the replacement and pick a font close to the original. Arial, Times New Roman and Helvetica cover most typed documents.
4. Match the size by eye, zooming in to compare letter height with the surrounding line.
5. Click Save and download the edited PDF.

This keeps the page looking like a scan, stamps and signatures included. The weakness is that the new text is typed while the rest is a photo, so a sharp-eyed reader may spot the difference. It also does not delete what is underneath. The original pixels remain in the file under the white box, so never use whiteout to hide private information. Redact PDF is the tool for that, and our guide on how to redact a PDF properly explains why.

Method 2: OCR, then edit in Word

When there is real writing to do, convert the scan into an editable document.

1. Straighten the scan first. If the pages are tilted, run Deskew PDF. If any are sideways or upside down, fix them with Rotate PDF. Recognition works best on level, upright lines.
2. Run OCR. Upload the file to OCR PDF and choose the main language printed on the page. Siliph offers English, Hindi, Spanish, French, German, Portuguese, Chinese (Simplified), Arabic and Japanese. Picking the wrong language is the most common cause of garbled output.
3. Choose the output. DOCX gives you the recognised text in a Word file as plain paragraphs, without the page layout. If you want the layout kept as closely as possible, download the Searchable PDF instead and convert it with PDF to Word, which rebuilds paragraphs, headings, tables and images.
4. Proofread, then edit. Fix recognition errors before making your own changes, so you do not confuse the two.
5. Export back to PDF. Use Word to PDF or Word's own Save as PDF.

The result is a clean, text-based document. It will not look like the original scan, and signatures or stamps on the paper come through as images at best. For a contract that will be re-signed anyway, that is usually fine.

Free alternatives

• Google Docs: in Google Drive, right-click the file and choose Open with, then Google Docs. Google's help page sets clear limits: "The file should be 2 MB or smaller", text "should be at least 10 pixels high", and documents "must be right-side up". It says bold, italics and font size are likely to be kept, but "Lists, tables, columns, footnotes, and endnotes are not likely to be detected" (Google Drive Help).
• Microsoft Word: Word can open a PDF and attempt to import its text. Microsoft notes that "The accuracy of the text recognition depends on the quality of the scan" and that "Handwritten text is seldom recognized" (Microsoft Support). It also warns that a page made mostly of graphics "might show up as an image", and then "the text can't be edited" (Microsoft Support).
• Adobe Acrobat: when you choose Edit PDF on a scan, "Acrobat automatically applies optical character recognition (OCR) to your document" (Adobe). Turning scans into editable PDFs is listed as an Acrobat Pro feature, which is a paid plan (Adobe plan comparison).

How to get cleaner OCR results

Most OCR errors come from the scan, not the software. A few habits help:

• Resolution: the Tesseract OCR project says it "works best on images which have a DPI of at least 300 dpi" (Tesseract documentation). If you are rescanning, use 300 dpi.
• Contrast: faint photocopies and grey paper produce misreads. Scan to PDF has a Boost contrast option for phone captures.
• Angle and shadows: hold the phone straight above the page and keep your shadow off it.
• What to check afterwards: names, numbers and anything with similar shapes, such as 0 and O, 1 and l, rn and m.

Handwriting is the hard limit. Printed text recognises well, but joined-up handwriting usually does not, and it will stay as an image.

Where your file goes

PDF Editor, OCR PDF, PDF to Word and Word to PDF all run on Siliph's servers. Files travel over an encrypted connection, are processed automatically with no human access, and are deleted within 2 hours. There is no account and no charge. Keep the original scan until you have checked the edited version.

One caution that has nothing to do with software: only edit documents you are entitled to change. Altering a certificate, a signed agreement or an official letter that someone else relies on can have legal consequences, however easy the tools make it.

Frequently asked questions

Why can I not edit the text in my PDF?

The PDF is almost certainly a scan, which stores each page as an image. There are no characters to select. Run OCR to create a text layer, or cover the old words and type new ones on top for a small fix.

Can I edit a scanned PDF for free?

Yes. Siliph's PDF Editor, OCR PDF and PDF to Word are free with no account. Google Docs can also convert scans up to 2 MB into editable text, though it drops tables, columns and lists.

Will OCR keep the original fonts and layout?

Not exactly. OCR recognises characters, not the typeface. A searchable PDF keeps the scan's appearance with hidden text behind it, while Word output gives editable text in standard fonts that you may need to restyle.

Can OCR read handwriting?

Rarely. OCR is designed for printed text. Neat block capitals are sometimes recognised, but cursive handwriting usually stays as an image that you cannot edit as text.

Is it safe to cover sensitive text with a white box?

No. The white box hides the words on screen and in print, but the original image is still in the file underneath. Use a redaction tool that removes the content, then check the result.