A scanned document looks like text but behaves like a picture. You can see the words, yet you can't select, search or edit a single one, because a scan is just an image of the page. OCR is what bridges that gap, turning a flat scan of a contract, form or letter back into editable text you can actually work with. This guide shows you how, and how to get a clean result the first time.
Why you can't edit a scanned document
When you scan paper, the scanner captures an image, a grid of pixels, not the underlying characters. That's why a scanned PDF often won't let you highlight text or run a find. The words are there visually, but there's no text layer behind them. OCR (optical character recognition) reads the image, identifies each character, and produces real, selectable text. If you'd like the concept explained from scratch, our what OCR is walkthrough covers it.
This matters for everyday paperwork: amending a contract, copying a clause, searching a stack of scanned letters, or reusing the text of a form you can no longer find the original of.
How to convert a scanned document to text
The exact tool depends on your file format.
If your scan is a PDF
Most scanned documents are saved as PDFs, often multi-page. Use the PDF to text tool:
- Upload the PDF. Drag the file into the converter, no install needed.
- Let the OCR layer run. For image-only, scanned pages, the tool automatically applies OCR rather than trying to copy a text layer that isn't there.
- Review the extracted text. It appears in reading order, ready to skim.
- Download. Save as TXT for plain text, or as DOCX if you want a Word-ready file.
If you need to keep formatting, headings, paragraphs and structure so you can edit the document like the original, use PDF to word instead, which rebuilds an editable .docx.
If your scan is an image
Single-page scans saved as JPG or PNG go through the image to text converter the same way: upload, let it read, then copy or download. The principles in our extract text from a photo guide apply here too.
Get a clean result: scanning tips
A scan's quality decides the outcome. These habits make a real difference:
- Scan at 300 DPI or higher. This is the sweet spot for printed text. Below about 150 DPI, fine character detail is lost and errors climb.
- Use black-and-white or greyscale for plain text. It gives the high contrast OCR relies on, and keeps files smaller.
- Keep the page straight. A skewed scan distorts the geometry the engine uses to find lines and words. Most engines deskew automatically, but feeding them a straight page helps.
- Flatten the page. For thick documents or books, press the page flat so the centre near the spine doesn't curve and blur.
- Clean the glass. Specks and smudges on a flatbed scanner show up as noise the engine has to work around.
Good scans of clean printed text convert with very high accuracy. Faint, old or skewed scans are still workable but warrant a closer proofread.
After conversion: what to check
Even strong results deserve a quick review, especially for paperwork where precision counts:
- Numbers and dates. Figures lack the dictionary context that auto-corrects words, so misreads here are most likely to survive.
- Names and signatures. Proper nouns aren't in any dictionary; check them.
- Tables and forms. If your scan is mostly tabular, you may get better structure from a format-aware tool, and for spreadsheets specifically, PDF to excel extracts rows and columns rather than flattening them into prose.
Once cleaned up, your scanned text is fully editable, searchable and reusable. Browse the full set of converters on our tools page to find the right output for your document.
Common scanned-document scenarios
The same workflow covers most of the paperwork people need to digitise:
- Contracts and agreements. Extract a clause to quote, or rebuild the whole document as an editable Word file so you can mark up revisions instead of retyping.
- Forms and applications. Pull the filled-in details into text, though heavily handwritten forms will be best-effort and may need correction.
- Letters and correspondence. Turn a scanned letter into searchable text so it's findable later, rather than a dead image in a folder.
- Receipts and statements. When the content is tabular, route it through PDF to excel so figures land in rows and columns instead of a flat block of text.
- Books and archived pages. Capture a passage from a scanned page without transcribing it by hand, the same approach used for extracting text from a photo of a printed page.
In each case the principle is identical: the scan is an image, OCR turns it back into usable text, and the quality of the scan decides how much proofreading you'll do afterward.
Frequently asked questions
Why can't I select text in my scanned PDF?
Because a scan is an image of the page, not actual text. There's no text layer behind the pixels for your PDF reader to select. Running the file through PDF to text applies OCR, which reads the image and produces real, selectable, editable text.
What's the best resolution for scanning documents for OCR?
Around 300 DPI is ideal for printed text. It captures enough character detail for reliable recognition without creating needlessly huge files. Scanning much lower than 150 DPI tends to lose the fine detail OCR needs and raises the error rate.
Can I keep the original formatting?
Yes, if you use a format-aware tool. PDF to word rebuilds an editable .docx that preserves headings, paragraphs and layout, rather than returning plain text. Choose plain text extraction when you only need the words, and the Word route when you need to edit the document like the original.
How accurate is OCR on scanned documents?
Clean scans of printed text at a good resolution convert with very high accuracy. Accuracy drops on faint, yellowed, skewed or low-resolution scans, so rescanning at higher DPI with good contrast is the most reliable way to improve a poor result. Always proofread numbers and names.
Have a scanned contract, form or letter to make editable? Start with PDF to text for documents, or image to text for single-page scans.