How to OCR a PDF and Make Scanned Documents Searchable

You open a PDF, press Ctrl+F, and search for a word you can clearly see on the page. Nothing happens. You try selecting a sentence, but your cursor cannot highlight the individual words.

The PDF probably contains scanned images rather than usable text.

Optical character recognition (OCR) can solve this problem by identifying text in scanned pages. With the right output settings, it can also add a searchable text layer while preserving the original page appearance.

This guide explains how to OCR a PDF using Adobe Acrobat, free local software, and online services. You'll also learn how to check whether OCR worked, improve recognition quality, and fix common problems.

Quick Answer: How to OCR a PDF

To make a scanned PDF searchable:

1. Open the PDF in a tool that supports OCR.

2. Choose the OCR or text recognition function.

3. Select the document language, if the tool offers that option.

4. Run OCR on the scanned pages.

5. Save the result as a searchable PDF, rather than exporting only plain text.

6. Open the saved PDF and test it by searching for a word and copying a short sentence.

7. Verify important information, especially names, dates, numbers, and tables.

If you already have Adobe Acrobat with OCR functionality, you can use its built-in text recognition workflow.

If you prefer a free local solution and are comfortable with command-line tools, OCRmyPDF is a strong option to evaluate.

For occasional non-sensitive documents, an online OCR service may be more convenient, although its privacy policy and file restrictions should be checked first.

The best method depends on your operating system, budget, document sensitivity, and the type of output you need.

How to Tell Whether Your PDF Needs OCR

Not every PDF requires OCR.

Some PDFs already contain selectable text. Others consist entirely of page images. A third category contains a mixture of searchable text and scanned pages.

Understanding the difference can save time and prevent unnecessary processing.

Test 1: Try Selecting a Sentence

Open the PDF in a reader and drag your cursor across a sentence.

If you can select individual words and copy them into a text editor, the page probably contains a text layer.

If you can select only the entire page as an image, the page may need OCR.

However, successful text selection does not guarantee that the underlying text is accurate.

A PDF may contain a defective or incomplete text layer, especially if it has already been processed by OCR software.

Test 2: Search for a Visible Word

Press Ctrl+F on Windows or Command+F on macOS.

Search for a distinctive word that clearly appears on the page.

If the search finds the correct word, the PDF likely has usable searchable text in that location.

If the word is visible but cannot be found, possible explanations include:

• The page is a scanned image.

• OCR was never performed.

• The existing OCR text contains recognition errors.

• The text uses an unusual encoding.

• The PDF reader is not interpreting the text layer correctly.

Try searching for several words on different pages before drawing a conclusion.

Test 3: Copy and Inspect the Text

Copy a paragraph into a plain-text editor.

Check whether:

• Words appear in the correct order.

• Spaces and punctuation are preserved.

• Numbers match the original.

• Special characters are readable.

• Multi-column content follows a sensible reading order.

For example, a page may visually display:

Total Revenue: $12,450.00

But a defective text layer could produce:

Total Revenue: $I2,45O.OO

The capital letter I has replaced the number 1, and the letter O has replaced zero.

This kind of error is particularly important when the PDF contains financial, legal, scientific, or operational information.

PDF Types and Recommended Actions

PDF Type	Typical Behavior	Recommended Action
Born-digital PDF	Text can usually be selected and searched	Use existing text; OCR usually unnecessary
Scanned PDF	Pages behave like images	Run OCR
Mixed PDF	Some pages are searchable; others are not	Use a tool that can handle mixed content
Previously OCRed PDF	Search works, but copied text may be wrong	Inspect and correct the existing text layer
Image-heavy PDF	Text, charts, and diagrams are mixed together	OCR text and inspect visual information separately

Important: OCR is not a general repair method for every PDF problem. If a document already contains accurate text, unnecessary OCR may add processing time or create new errors.

How to OCR a PDF With Adobe Acrobat

Adobe Acrobat provides an OCR workflow for scanned documents.

The exact interface and feature availability may depend on the Acrobat product, version, and subscription.

Step 1: Open Your Scanned PDF

Launch an Acrobat version that includes OCR functionality.

Open the scanned document and confirm that its pages display correctly.

If the PDF is important, save a backup copy before processing it.

Step 2: Open the OCR Function

In supported current Acrobat desktop interfaces, the workflow is generally:

All tools → Scan & OCR → In this file

Choose the pages you want to process.

If the document contains several languages, review the recognition settings before proceeding.

Adobe may change interface labels or feature placement between releases, so consult the current product documentation if your menus differ.

Step 3: Recognize the Text

Start the text recognition process.

Processing time depends on the document's length, scan quality, and system resources.

Do not assume that the absence of an error message means every word was recognized correctly.

Step 4: Save the Searchable PDF

Save the processed document under a new filename, such as:

annual-report-searchable.pdf

Using a new filename makes it easier to compare the original scan with the OCR result.

Step 5: Verify the Result

Close and reopen the saved PDF.

Then:

1. Search for a distinctive word on the first page.

2. Search for another word on a later page.

3. Copy a paragraph into a text editor.

4. Check a number, date, or name against the original scan.

5. Confirm that the page layout remains suitable for your intended use.

If these checks fail, inspect the recognition settings and source scan before repeating the process.

When to choose Acrobat: It is worth considering when you already use Acrobat for PDF editing and document management, and its available OCR features meet your needs.

For a broader comparison of PDF software, see our guide to the best PDF software.

How to Correct OCR Recognition Errors in Adobe Acrobat

If Acrobat recognizes most of the document but makes mistakes in individual words, you may not need to repeat OCR on the entire file.

In supported Acrobat desktop versions:

1. Open the processed PDF.

2. Select All tools → Scan & OCR.

3. Choose Correct recognized text.

4. Enable Review recognized text.

5. Select a highlighted word that Acrobat has marked for review.

6. Compare the recognized text with the original scanned image.

7. Correct the text where necessary and accept the change.

8. Save the corrected PDF under an appropriate filename.

This is particularly useful when OCR confuses similar-looking characters in names, dates, or numbers.

However, automated review indicators may not identify every error. Manually inspect information that is important to your task.

For the current procedure, consult Adobe's guide to correcting recognized text.

How to OCR a PDF for Free

You do not necessarily need a paid PDF editor to perform OCR.

Free options include open-source local software and some online services.

However, "free OCR" can mean different things. A service may offer a limited number of pages, restrict file size, require an account, or export only text rather than a searchable PDF.

Check what the tool actually produces before relying on it.

Method 1: Use OCRmyPDF for Local OCR

OCRmyPDF is an open-source command-line application designed to add OCR text to scanned PDF files.

It uses OCR technology, including Tesseract, and supports workflows that preserve the original scanned page appearance while adding searchable text.

This makes it useful for users who want more control over local document processing.

Step 1: Install OCRmyPDF

Follow the installation instructions for your operating system in the official OCRmyPDF documentation.

Installation requirements differ across Windows, macOS, and Linux.

Depending on the platform and installation method, you may also need OCR engines, language data, or supporting utilities.

Do not assume that the same installation command works on every system.

Step 2: Prepare Your Input File

Place the scanned PDF in a working folder.

For example:

scanned-report.pdf

Keep the original document unchanged.

Step 3: Run a Basic OCR Command

Open a terminal in the folder containing the file and run:

ocrmypdf scanned-report.pdf searchable-report.pdf

The first filename is the input document.

The second filename is the output document.

This command is a basic documented usage pattern, not a guarantee that every PDF can be processed without additional options.

Step 4: Specify a Recognition Language

For an English-language document, you can explicitly specify English:

ocrmypdf -l eng scanned-report.pdf searchable-report.pdf

The relevant Tesseract language data must be installed.

For documents containing more than one language, OCRmyPDF supports multiple language specifications when the required language data is available.

Step 5: Handle PDFs That Already Contain Text

Some PDFs contain a mixture of scanned pages and pages with existing text.

In OCRmyPDF 17.0.0 and later, you can use:

ocrmypdf --mode skip scanned-report.pdf searchable-report.pdf

This tells OCRmyPDF to skip OCR on pages that already contain text.

For compatibility with older workflows, the following option is also supported:

ocrmypdf --skip-text scanned-report.pdf searchable-report.pdf

Be careful: skipping a page with existing text does not repair an inaccurate OCR layer.

If the existing OCR text is defective, consult the documentation for --mode redo. Avoid forcing OCR on every page without understanding the consequences, because force mode rasterizes existing content.

Always preserve the original file and verify the output before using it for important work.

See the OCRmyPDF processing modes documentation for the current options.

Step 6: Inspect the Output

Open searchable-report.pdf in a PDF reader.

Search for words on several pages and copy representative text.

Check that the page images remain readable and that important information has not been lost.

When to choose OCRmyPDF: It is particularly useful when you want a local, scriptable workflow and are comfortable installing and operating command-line software.

Limitations: Installation can be more technical than a graphical OCR tool. Recognition quality still depends on the source document, languages, layout, and OCR engine.

Method 2: Use an Online OCR Service

Online OCR tools can be convenient when you need to process a small document without installing software.

A typical workflow is:

1. Open a reputable OCR service.

2. Review its privacy policy and supported file formats.

3. Upload a non-sensitive scanned PDF.

4. Select the recognition language if available.

5. Choose the desired output format.

6. Run OCR.

7. Download the result.

8. Check whether the output is actually a searchable PDF.

Some services return a Word document or plain text rather than a PDF with a searchable text layer.

That distinction matters if you need to preserve the original document appearance.

Privacy warning: Avoid uploading confidential contracts, private medical records, financial statements, customer data, or unpublished research to a third-party service unless the upload is authorized and the provider's data handling meets your requirements.

For sensitive files, an appropriately configured local OCR workflow may be preferable.

Which OCR Method Should You Choose?

No single OCR tool is ideal for every document.

Use the following decision table to narrow your options.

Your Situation	Method to Consider	What to Check
You already have Acrobat with OCR	Adobe Acrobat	OCR availability in your version
You want a free local workflow	OCRmyPDF	Installation and language support
You have a small, non-sensitive document	Online OCR service	Privacy, limits, output format
You need to process many files	Scriptable local OCR	Automation and error handling
Your PDF contains tables	OCR followed by table extraction	Whether rows and columns survive
Your document contains confidential information	Approved local or enterprise workflow	Data handling and permissions

Choose a method based on the required result, not just how quickly it promises to recognize text.

For example, extracting words from a scanned invoice is not the same as producing an accurate spreadsheet containing invoice line items.

How to Make a Scanned PDF Searchable Without Changing Its Appearance

One of the most useful OCR outputs is a searchable PDF that retains the original scanned page image.

This typically works by placing recognized text in a text layer associated with the scanned page.

The visible page can continue to look like the original scan, while the text layer allows search and selection.

Searchable PDF vs Plain Text vs Editable Document

Output	Main Purpose	Important Limitation
Searchable PDF	Preserve page appearance and enable text search	OCR text may still contain errors
Plain text	Extract words for further processing	Formatting and layout are usually lost
Word document	Make text easier to edit	Original layout may change
CSV or Excel	Work with structured data	Requires correct row and column extraction

Suppose you have a scanned research report with 40 pages.

If your goal is to find specific terms and quote passages, a searchable PDF may be enough.

If you need to revise the report's wording, an editable document might be more appropriate.

If you need to analyze its financial tables, OCR is only one part of the process.

Does OCR Make Every PDF Fully Editable?

No.

A searchable PDF is not automatically a fully editable document.

OCR identifies text, but it does not necessarily reconstruct the original fonts, paragraphs, table structure, images, or document layout as editable elements.

Some PDF editors offer additional conversion or editing functions, but their results should be checked separately.

How to Improve OCR Accuracy

OCR accuracy depends heavily on the source document.

A clear, properly oriented scan is usually easier to recognize than a blurred, skewed, low-contrast image.

The following improvements can help.

1. Use the Best Available Source

Whenever possible, start with the original digital document rather than a scan.

If only a scan is available, use the highest-quality original file you can access.

Avoid repeatedly converting or compressing the same document before OCR.

2. Correct Page Rotation and Skew

A page that is sideways or significantly tilted can make recognition harder.

Some OCR tools provide automatic rotation or deskewing options.

For example, OCRmyPDF supports image-processing options that may help with rotated or skewed scans.

These transformations can affect the page image, so preserve the original and inspect the output.

3. Choose the Correct Language

OCR engines rely on language models or recognition data.

Selecting the wrong language can reduce recognition quality, especially for accented characters or non-English scripts.

For multilingual documents, check whether the selected tool supports all required languages.

4. Watch for Similar-Looking Characters

Common OCR confusions include:

Intended Character	Possible Recognition Error
0	O
1	I or l
5	S
8	B
rn	m

These are illustrative examples of possible recognition errors, not results from a controlled test.

Pay special attention to account numbers, dates, invoice totals, measurements, and technical identifiers.

5. Treat Tables as a Separate Challenge

OCR can recognize individual words and numbers without understanding their relationship to rows and columns.

For example, a table containing three product names and three prices may be recognized as six separate text fragments.

The text might be accurate while the data structure is wrong.

If you need spreadsheet-ready output, follow our guide to extracting tables from PDFs with AI.

6. Check Multi-Column Documents

Research papers, magazines, and reports often use multiple columns.

OCR may recognize the words correctly but produce an incorrect reading order.

For example, it might combine the first line of the left column with the first line of the right column.

Check paragraph continuity before using extracted text for summarization or analysis.

7. Do Not Assume Handwriting Will Be Recognized Reliably

Handwriting recognition is a different and often more challenging task than recognizing printed text.

Results depend on the writing style, image quality, and capabilities of the recognition system.

If handwriting is central to your document, choose a tool designed for that task and verify the output carefully.

How to Check Whether OCR Worked Correctly

An OCR process is complete only when the resulting document meets your intended requirements.

A successful file export is not enough.

Use the following checks.

Check 1: Searchability

Search for several distinctive words across different pages.

Include a word near the beginning, middle, and end of the document.

Check 2: Text Selection

Select a complete sentence and copy it into a plain-text editor.

Confirm that the copied words match the scanned page.

Check 3: Numerical Accuracy

Inspect important numbers such as:

• Prices and totals.

• Dates and years.

• Percentages.

• Measurements.

• Product codes.

• Reference numbers.

For documents where numerical accuracy is critical, check the original scan rather than relying on OCR text alone.

Check 4: Reading Order

Copy a paragraph from a multi-column or complex page.

Confirm that the text follows the intended sequence.

Check 5: Page Coverage

Check whether every required page was processed.

A document can appear searchable while still containing unrecognized pages.

Check 6: Layout Preservation

Compare the processed PDF with the original.

Look for missing pages, altered orientation, cropped content, or unexpected visual changes.

A Practical OCR Verification Checklist

Check	Pass Condition
Search	Known words can be found on representative pages
Copy	Selected text matches the visible scan
Numbers	Critical values match the original
Reading order	Paragraphs follow a logical sequence
Coverage	All required pages contain usable text
Appearance	The PDF remains suitable for its intended purpose

For high-stakes documents, this checklist is a starting point, not a substitute for a more rigorous review.

Practical Example: Convert a Scanned Report Into a Searchable PDF

Consider a hypothetical 12-page scanned business report.

The file contains printed English text, section headings, and a financial table.

The goal is to create a searchable PDF for internal reference without changing the original page appearance.

This is an illustrative workflow, not a claim that a specific document was independently tested.

Input

business-report-scan.pdf

Assumptions:

• The file contains scanned page images.

• Text cannot currently be selected.

• The pages are reasonably legible.

• The user has permission to process the document.

Step 1: Preserve the Original

Keep the source PDF unchanged.

Step 2: Run Local OCR

With OCRmyPDF correctly installed, use:

ocrmypdf -l eng business-report-scan.pdf business-report-searchable.pdf

Step 3: Inspect the Saved PDF

Open the output and search for a section heading that appears on page 2.

Repeat with a distinctive word on page 10.

Step 4: Copy a Paragraph

Copy a paragraph from the middle of the report.

Check its punctuation, spacing, and reading order against the scanned page.

Step 5: Verify the Financial Table

Compare the visible table values with any text extracted from that page.

Do not assume that recognized text preserves the table's original rows and columns.

Step 6: Decide Whether the Output Is Fit for Purpose

If the report is now searchable and its important text is correct, the OCR process may be sufficient for document navigation and basic reading.

If the goal is to analyze the financial table, continue with a separate table extraction and validation workflow.

If the output contains errors, correct the scan or OCR settings and repeat the verification process.

The important distinction is between OCR completed and document ready for its intended use.

Common OCR Problems and How to Fix Them

Problem 1: The PDF Is Still Not Searchable

Possible causes:

• The OCR process did not finish.

• The wrong output format was selected.

• Some pages were skipped.

• The recognized text was not saved into the PDF.

• The PDF reader is having difficulty interpreting the output.

What to do:

Open the saved file in another PDF reader, test multiple pages, and confirm that the tool produced a searchable PDF rather than an image-only copy.

Problem 2: OCR Produces Garbled Text

Possible causes:

• Low-resolution scanning.

• Blurred or compressed images.

• Incorrect language settings.

• Unusual fonts.

• Poor contrast.

What to do:

Use a better-quality source, correct page orientation, check language settings, and test a representative page before processing the entire document again.

Problem 3: Text Appears in the Wrong Order

Possible causes:

• Multiple columns.

• Complex page layouts.

• Tables.

• Sidebars or footnotes.

What to do:

Inspect the page visually and compare the copied text with the original reading order.

For complex documents, consider a layout-aware extraction method rather than relying on basic OCR output alone.

Problem 4: The Tool Says the PDF Already Contains Text

The PDF may already contain a text layer.

First determine whether that text is accurate.

If it is accurate, OCR may not be necessary.

If it is incorrect, review the tool's documented options for handling existing OCR or text layers.

Do not remove or replace an existing text layer without preserving the original document.

Problem 5: OCR Works on Text but Fails on Tables

OCR and table extraction are related but separate tasks.

OCR may identify the text inside a table while losing the relationships between cells.

Use a structured table extraction workflow and compare its output with the original table.

Problem 6: The File Is Too Large for an Online OCR Service

Online tools may impose file size, page count, or usage limits.

For large documents, consider a desktop or local solution that supports your workflow.

Avoid splitting a document into multiple files without a plan for preserving page order and verifying the final output.

OCR vs AI PDF Tools: What Happens After Text Recognition?

OCR makes scanned text available for searching or further processing.

AI PDF tools can help perform additional tasks, such as summarizing, explaining, comparing, and extracting information.

But these are different stages of a document workflow.

Task	Primary Function
OCR	Recognize text from scanned images
PDF text extraction	Retrieve text already present in a PDF
AI summarization	Produce a shorter explanation of document content
Document comparison	Identify similarities and differences
Table extraction	Recover structured rows and columns
Source verification	Check whether claims match the original material

A searchable PDF may be easier for some AI workflows to process, but OCR does not guarantee that an AI assistant will interpret the document correctly.

For example, the OCR text may contain a mistaken number, or the AI may misunderstand a chart that requires visual interpretation.

This is why the original PDF should remain available for verification.

For more detail, read Why ChatGPT Misreads PDFs and How to Fix It.

If you want to summarize or analyze a document after OCR, our comparison of AI tools for PDFs explains the different tool categories.

Frequently Asked Questions

Can I OCR a PDF for free?

Yes. Free local options such as OCRmyPDF are available. Some online services also provide free OCR, although their limits and output formats vary.

Can I make a scanned PDF searchable without changing its appearance?

Often, yes. OCR software can add a searchable text layer while retaining the scanned page image. The resulting file should still be inspected for recognition and layout issues.

Is a searchable PDF the same as an editable PDF?

No. A searchable PDF allows text searching and often copying, but it does not necessarily provide fully editable paragraphs, tables, or page layouts.

Why does Ctrl+F fail in a scanned PDF?

The visible words may exist only as pixels in a page image. Without a usable text layer, the PDF reader has no searchable words to match.

Does OCR work on handwritten PDFs?

Some recognition tools support handwriting, but performance varies considerably. Handwritten documents should be checked carefully, especially when the content is important.

Can OCR extract tables into Excel?

OCR can recognize text within tables, but it may not preserve the structure. Accurate spreadsheet conversion usually requires a separate table extraction and validation step.

Is online PDF OCR safe for confidential documents?

It depends on the service, its data handling practices, your organization's requirements, and the sensitivity of the document. Do not upload confidential files without appropriate authorization.

Can ChatGPT OCR a scanned PDF?

ChatGPT may be able to interpret some scanned or image-based content, depending on the available features, file handling, and document characteristics. However, that is not the same as reliably creating a searchable PDF with a preserved text layer.

For a reusable searchable PDF, a dedicated OCR workflow is often more appropriate.

Conclusion: Choose OCR Based on the Result You Need

Learning how to OCR a PDF is not just about making visible words searchable.

The goal is to produce a document that is useful and sufficiently accurate for the task that follows.

If you already have suitable Adobe Acrobat functionality, its integrated OCR workflow may be convenient.

If you want a free local approach and can work with command-line tools, OCRmyPDF is worth considering.

If you only need to process an occasional non-sensitive document, an online OCR service may be sufficient.

Whichever method you choose, verify the output. Test searchability, copied text, important numbers, reading order, and page coverage.

And remember that OCR is only the first step when your ultimate goal is document analysis, table extraction, or AI-assisted research.

The best OCR result is not merely a file that can be searched. It is a document whose recognized information you can check and use with confidence.