Make scanned PDFs searchable with OCR
Turn scanned PDFs into searchable, copyable documents with OCR. Learn how to scan well, pick languages, handle mixed text and check the result for errors.
A folder of scanned PDFs looks organized until you need to find something inside it. You remember a supplier invoice mentioned a particular part number, but searching does nothing, because every page is just a picture. You end up opening files one by one and squinting.
OCR (optical character recognition) fixes this. It reads the text in a scanned image and adds it to the PDF as real, searchable text. This guide explains how OCR works, how to get accurate results, and what to do with your documents once they are searchable.
What OCR actually does
When you scan a sheet of paper, the scanner takes a photograph. The resulting PDF contains an image of the page, but no text. To a computer, the word "Invoice" at the top is just a pattern of dark and light pixels.
OCR analyzes those pixels, recognizes letters and words, and adds an invisible text layer to the PDF that lines up with the image. The page still looks exactly the same, but now you can:
- Search for words across the document
- Select and copy text
- Have assistive tools such as screen readers read it aloud
- Use the text in other tools, such as conversion, summaries or data extraction
How to tell if a PDF needs OCR
Open the PDF and try to select a single word. If the word highlights on its own, the PDF already has text. If you can only select the whole page, or nothing at all, it is an image and needs OCR. Searching for a word you can clearly see on the page is another quick test.
Running OCR step by step
PDFbat includes OCR in 45 languages, with the option to add a second language for mixed documents.
- Open the scanned PDF in PDFbat, from your device or your PDF Drive.
- Choose the OCR tool.
- Pick the main language of the document.
- Add a second language if the document mixes two, such as a bilingual contract.
- Run OCR. The result is saved to your PDF Drive automatically.
- Test it. Search for a few words you can see on the page and copy a paragraph to check accuracy.
Mixed language documents
Many real documents use more than one language: a residence permit in the local language with English headings, a product manual in two languages side by side, or a letter with a quoted passage. Setting both languages lets OCR recognize each one properly rather than forcing everything through one language's rules.
Getting accurate results starts at the scanner
OCR can only read what the scan shows clearly. A few habits make a big difference.
Scan settings
- Resolution around 300 DPI is a good standard for text documents. Much lower, and small print becomes hard to read. Much higher mostly makes files bigger.
- Grayscale or black and white is fine for most text. Use color only when the original has colored content that matters.
- Flat pages. Press books and folded letters flat so lines are straight.
Phone scans
If you are using a phone camera:
- Place the page on a dark, plain surface so the edges stand out.
- Use even light and avoid shadows from your hand or phone.
- Hold the phone directly above the page, not at an angle.
- Fill the frame with the page.
Fix the pages first
Before running OCR, use PDFbat's other tools to clean up:
- Rotate pages that came in sideways or upside down.
- Delete blank pages and duplicates.
- Crop away large dark borders.
- Repair a file that will not open properly.
Our guide on merging, splitting and organizing PDFs covers these tools in more detail.
Checking the result
OCR is very good on clean, printed text, but it is not perfect, and it is worth knowing where it struggles.
| Content | What to expect |
|---|---|
| Clean printed text | Usually reads well |
| Small print, faint photocopies | More mistakes, check carefully |
| Tables | Text is captured, but check numbers in each cell |
| Handwriting | Often unreliable |
| Stamps, logos, decorative fonts | Often misread or skipped |
| Skewed or curved pages | Accuracy drops toward the edges |
Watch the numbers
Letters can be guessed from context, but numbers cannot. A "5" read as "S", or a "0" read as "O", is easy to miss and can matter a lot on an invoice or a contract. If you plan to rely on figures, spot check them against the image.
What you can do once a PDF is searchable
Searchable text opens up a lot more than search.
Edit and convert
With a text layer in place, you can copy passages into other documents or convert the PDF to an editable Word file. See how to convert a PDF to Word for that process.
Redact properly
PDFbat's true redaction deletes the text, images and vector art under the redaction box, rather than just covering it. With OCR text in the file, there is real text to remove, so redact after running OCR when you need to share a scanned document that contains personal details.
Ask questions with AI
PDFbat AI can chat with a PDF and answer with page citations, including scanned documents, for files up to 300 pages. It can also summarize, translate a whole PDF, or extract data to JSON or CSV, such as pulling dates and totals from a stack of scanned receipts. AI features use AI credits from your OfficeBats wallet. Read AI credits explained to see how credits work.
Keep file sizes manageable
Scans are large. After OCR, you can run Compress to bring the size down. Keep the text sharp by choosing a lighter level and checking small print afterward. Our guide to compressing a PDF without losing quality walks through the settings.
A quick example
A bookkeeper for a small contractor has a few years of scanned supplier invoices. They rotate and clean up each batch, run OCR in English, and file everything in PDF Drive folders by year. Now, when a question comes up about a particular part number or supplier, a search finds the right invoice in seconds instead of an afternoon of scrolling.
Limits worth knowing
OCR is one of PDFbat's tools. On the free plan you get 10 tool tasks a day with files up to 25 MB. Large scanned archives can hit that file size quickly, so Pro and Business raise the limit to 100 MB with unlimited tasks. Compare plans on the pricing page.
The short version
- Scanned PDFs are images. OCR adds a searchable text layer without changing how pages look.
- Scan at around 300 DPI, flat, well lit and straight.
- Pick the right language, and add a second one for mixed documents.
- Clean up pages before OCR, and spot check numbers after.
- Once searchable, you can convert, redact, compress and ask AI about your scans.
Open PDFbat and try OCR on the scan you have been meaning to search.
