How to Edit a Scanned PDF: 3 Methods That Actually Work
A scanned PDF is a picture of a page, so it cannot be edited like normal text. Here are three ways around that, and how to pick one.
On this page
- The quick answer
- Method 1: add text on top of the scan
- Method 2: OCR, then edit the text
- Method 3: rebuild the document
- Getting accurate text from a scan
- A worked example: changing an address on a scanned letter
- Mistakes to avoid
- Common problems with scanned documents, and fixes
- When to ask for the original instead
- Protecting scanned documents
- A short checklist
A scanned PDF looks like a document, but each page is really a photograph of paper. There are no words in the file, only pixels, which is why you cannot select the text, search it, or change it the way you would in a normal document. That does not make a scan uneditable. There are three practical ways to deal with one, and the right choice depends on how much needs to change and how polished the result has to be.
The quick answer
- Add on top: place text, a date, a check mark or a signature over the scanned page. Fastest, but the scan underneath is unchanged.
- OCR then edit: recognise the text, copy it into a word processor, and edit it there. Best for changing a lot of text.
- Rebuild: retype or recreate the document from scratch, using the scan as a reference. Best for a polished result.
Method 1: add text on top of the scan
If you only need to add something, such as a date, a name, an initial or a signature, you do not need to touch the scanned text at all. Open the PDF in an editor that lets you place text, shapes or a signature on the page, position it carefully, and save a new copy. This is the standard way to complete a form that arrived as a scan. Match the size of the printed text, keep each entry above its line, and zoom in to check alignment. Our guides to adding text to a PDF and to filling out and signing a PDF form cover the placement details.
Need to add text, a date or your signature to a scan? Place it right on the page.
Open Fill & Sign →Be aware of the limit. This does not change the scanned words. If a figure in the scan is wrong, covering it with a patch and typing over it is a visible repair, and for anything official it is better to correct the source or rescan. Also note that an editor that edits existing text lines works on real text, so it cannot rewrite the words in a picture.
Method 2: OCR, then edit the text
OCR, or optical character recognition, reads the pictures of letters and turns them into real characters. The usual route is to recognise the text, copy it into Word or another editor, and make your changes there. Our tools turn this into a few steps.
- Convert the scanned PDF pages to images, if your OCR tool works on images, naming them page-01, page-02 and so on.
- Run each image through image to text (OCR) and copy the recognised text.
- Paste it into Word, with a heading for each page, and tidy the line breaks.
- Proofread every number, date and name against the scan.
- Edit as needed, then export a new PDF.
The detailed steps and the common recognition errors are in how to extract text from a scanned PDF, and background on how recognition works is in OCR explained. Expect to reformat: OCR recovers words, not layout.
Method 3: rebuild the document
When the result has to look professional, or the scan is too poor for OCR, the cleanest route is to recreate the document. Retype it, or paste the recognised text into a fresh template, and use the scan only as a reference. It takes longer, but you end up with a real document: searchable, accessible, small, and easy to change next time. Keep the scan as a record of the original.
| Add on top | OCR then edit | Rebuild | |
|---|---|---|---|
| Best for | Dates, names, ticks, signatures | Changing a lot of text | A polished, reusable document |
| Speed | Fastest | Medium | Slowest |
| Result is searchable | No | Yes, in the new document | Yes |
| Looks like the original | Yes, with additions | No; layout is rebuilt | As good as you make it |
| Accuracy risk | Low | Recognition errors | Typing errors |
| Keeps the scan intact | Yes | Yes | Yes, kept as reference |
Getting accurate text from a scan
OCR accuracy depends mostly on the scan. Resolution, straightness, contrast and lighting matter far more than the software, and improving the scan is the cheapest way to improve the result.
- 100 DPI0.94 MP
- 150 DPI2.1 MP
- 300 DPI8.42 MP
- 600 DPI33.66 MP
Calculated from the page size of 8.5 by 11 inches. Around 300 DPI is a common sweet spot for printed text; much lower loses detail, much higher mostly adds file size.
- Scan at around 300 DPI for normal printed text.
- Keep pages flat and straight, with even lighting and no shadows.
- Use greyscale or black and white for plain text, to keep files small and contrast high.
- Crop away borders and rotate pages upright before recognition.
- Choose the correct language in the tool, if it has a setting, so accents are recognised.
- Process one clean column at a time for multi-column pages.
A worked example: changing an address on a scanned letter
You have a scanned two-page letter, and the recipient's address on page one is wrong. If you only need a copy for your own records, you cover the old address with a small patch and type the new one on top, matching the font size. If the letter will be sent as an official document, you instead recognise the text, paste it into a letter template, correct the address, and export a fresh PDF. You compare the new document with the scan, read the numbers carefully, and keep the scan on file. The first route took two minutes and is fine for a casual copy. The second took twenty minutes and produces a real document.
Mistakes to avoid
- Expecting to edit the scanned words directly.
- Trusting recognised text without proofreading numbers and names.
- Covering text and typing over it in an official document.
- Scanning at a very low resolution, then trying to fix it with software.
- Throwing away the original scan or paper.
For a wider look at what editing involves, read what a PDF editor is, and for how accurate free recognition tends to be, how accurate free OCR software is. If you would rather convert the whole file, see how to convert PDF to Word.
Common problems with scanned documents, and fixes
| Problem | What you see | Fix |
|---|---|---|
| Skewed page | Text slopes across the page | Rescan straight, or straighten the image before OCR |
| Shadows and uneven light | Dark patches, uneven background | Rescan in even light, or raise contrast |
| Show-through | Text from the other side appears faintly | Place black paper behind the page, or use a higher contrast setting |
| Low resolution | Letters look blocky or blurred | Rescan at about 300 DPI |
| Faded print | Text is pale grey | Rescan in greyscale with more contrast |
| Curved book pages | Distorted lines near the spine | Press the page flat or photograph the page in sections |
Fixing the scan is nearly always cheaper than fighting the consequences later. A minute spent rescanning a page can save ten minutes of correcting recognised text.
When to ask for the original instead
A scan is a copy of a copy, and it is the worst starting point for editing. If the document came from an organisation, a colleague or a website, ask whether they have the original digital file, or a text-based PDF. Many do, and they can export one in seconds. A digital original is more accurate, smaller, searchable and accessible, and it spares you every step in this guide. A short message asking for a text version is often the fastest edit you will ever make.
Protecting scanned documents
- Scans often contain personal data such as addresses, identity numbers and signatures, so store them somewhere secure.
- Think before uploading a scan to a service, and remove pages that are not needed.
- Add a password before sharing a sensitive scan, following how to password protect a PDF.
- Delete loose scan images from your phone or downloads folder once the PDF is finished.
- Keep the original paper or best-quality scan as the authoritative record.
A short checklist
- Check whether the PDF is really a scan.
- Choose add-on-top, OCR-then-edit, or rebuild based on how much needs to change.
- Improve the scan first if recognition is poor.
- Proofread names, numbers and dates against the original.
- Keep the scan, and save your edited version under a new name.
Frequently asked questions
Can you edit a scanned PDF?
Not directly, because each page is a picture. You can add text on top of it, or you can run text recognition (OCR) to get editable text and then edit that.
How do I make a scanned PDF editable?
Run OCR on the pages to recognise the text, then copy the text into a document such as Word and edit it there. Proofread carefully, since recognition makes mistakes.
Why can't I select text in my PDF?
It is probably a scan. A scan is an image of a page, so there is no real text to select until OCR adds a text layer.
How accurate is OCR on a scanned PDF?
On clean, straight, 300 DPI scans of printed text it is usually good, with a few errors per page. Faded, skewed, handwritten or complex pages are much worse. Always proofread.
What is the easiest way to change one word in a scan?
Cover the old word and place new text on top, matching the font size and colour. This is quick, but for anything official it is better to rescan or recreate the document.
Should I edit a scan or create a new document?
If you need to change a lot or the result must look professional, recreate it from recognised text. If you only need to add a date or a note, placing text on top is fine.