OmniDocOCR PDF to Text: Extract Text from Scanned Files
Turn a scanned PDF into plain, selectable text with OCR, and learn when to skip layout and export to Word or Excel instead.
A scanned PDF looks like text, but to a computer it is a picture of each page. If you try to select words or paste them somewhere, you get nothing, or a block of unusable characters. When you need the words themselves, for quoting a passage, searching a contract for a clause, or pasting into a notes file, you need OCR (optical character recognition). OCR reads the letters on each page and adds a text layer you can select and copy.
This guide is for anyone holding a scanned contract, a printed report or a photographed form who needs the plain text and nothing more. Plain text leaves out fonts, columns and table borders. That is exactly what you want when the goal is the content, and exactly what you lose when you need the page to look the same. If layout matters, the same OCR step can feed a Word or Excel export instead.
Below you will find a short answer, the checks to make before you start, the steps to run OCR in the OmniDoc desktop client and copy the text out, and a way to confirm the result is right.
How do you get plain text out of a scanned PDF? The short answer
Run OCR on the scanned PDF so each page gets a text layer, then copy the text you need or export it, and check the result against the original before you rely on it.
| Your situation | What to do | Why |
|---|---|---|
| You need the words only, for quoting or searching | Run OCR in the desktop PDF editor, then copy the text | Plain text carries no layout to rebuild |
| You need text you will keep editing in a document | Run OCR, then export to Word | Word keeps paragraphs and headings you can work with |
| You need numbers arranged in rows and columns | Run OCR, then export to Excel | Columns come through as cells instead of loose lines |
| The PDF already lets you select text | Copy directly and skip OCR | OCR adds nothing to a page that already has a text layer |
| Pages are blurry, skewed or handwritten | Rescan if you can, then run OCR and check closely | Misreads become more likely as characters get harder to read |
Before you start
- Confirm the PDF is scanned. Try to select a line of text. If nothing highlights, or the whole page selects as one image, the file is scanned. Running OCR on a page that already has text is unnecessary, and you can copy from it directly.
- Check the scan quality. Look at a few pages at full size. Small, faint or cut-off letters are where OCR most often goes wrong, so a flat, complete scan of each page gives you a better starting point.
- Decide the output format. Plain text, Word and Excel each keep a different amount of structure. Choosing now tells you whether to copy the text or export it.
- Install the desktop client. The PDF editor, which includes OCR, runs in the OmniDoc desktop client for Windows, macOS and Linux. You can download it at download.omnidoc.top.
- Keep the original file untouched. You will compare the extracted text against it, and an unchanged copy is your reference if something looks off.
OCR PDF to text in the desktop client, step by step
- Open the scanned PDF in the PDF editor. The editor is where OCR and the export options live, so start there rather than in a separate tool.
- Run OCR on the document. OCR reads each page image and adds a layer of text that can be selected. For a long file, you can limit the work to the pages you need.
- Check the first page before going further. Compare a paragraph on page one with the scan. If the first page is badly misread, the whole file likely needs a better scan.
- Select the text you need and copy it. After OCR, the text on each page can be selected like any digital PDF, so you can copy one paragraph or a whole page.
- Paste into a plain-text editor and remove the noise. Repeated page numbers, running headers and footers get copied along with the body text, so delete them before you save.
- Export to Word or Excel if you need structure. Plain copying gives you lines in reading order, which can merge columns or break tables. Export keeps paragraphs for Word and cells for Excel.
- Save the result under a name that points back to the source. A file name that includes the original title and date makes it clear later which scan the text came from.
Check the result
OCR is not a guarantee of exact characters, so check the output before you use it, especially for contracts, medical records or anything with figures.
| Check | How to see it | If it's wrong |
|---|---|---|
| Letters and digits | Compare a paragraph and a number line by line with the original | Correct the characters by hand, or rescan the page if many are wrong |
| Reading order and line breaks | Read the first and last lines of each page | Copy page by page, then rejoin; for multi-column pages, check the order carefully |
| Missing lines | Count paragraphs against the original page | Re-run OCR on that page and check the scan for cut-off edges |
| Stray text | Look for page numbers and repeated headers | Delete them before you save or paste the text anywhere |
| Dates, amounts and names | Check each one against the source | Treat them as unverified until someone has confirmed them |
Common misreads include similar-looking pairs such as "l" and "1", "O" and "0", or "rn" and "m". Faint print, unusual fonts and skewed pages make these more likely. Note also that handwriting is harder for OCR than printed text, so expect more errors there.
❓ FAQ
1. Can OCR read a photo or image file, not just a PDF?
The OCR step in the desktop PDF editor works on scanned PDFs. For a photo, put the image into a PDF first using a general-purpose converter, then open that PDF and run OCR on it.
2. Why does my extracted text have wrong letters or symbols?
OCR works from the shapes it sees, so small, faint, skewed or unusually styled characters are often misread. Rescanning at better quality usually helps more than editing afterward. For the few spots that remain, correct them by hand against the original.
3. Will my table or columns survive if I copy the text?
Usually not. Copying gives you lines in reading order, so columns can merge and table cells run together. If you need the rows and columns, export to Excel instead of copying.
4. Can I just upload a scanned PDF for translation instead?
Yes. OmniDoc translation accepts scanned PDFs as they are, and you can choose bilingual side-by-side output or translation-only output on omnidoc.top. Use translation when you need a translated document, and use OCR and copy when you need the original words as plain text.
📝 Summary
Getting plain text from a scanned PDF comes down to three steps: run OCR so the pages gain a text layer, copy or export the result, and check it against the original. Choose plain text when you only need the words, and Word or Excel when you need the structure back. Always verify names, numbers and dates before you rely on them.
