How to OCR a Scanned PDF and Make It Searchable

How to OCR a Scanned PDF and Make It Searchable

You open a PDF, try to select a sentence, and your cursor just drags a box over the page. Nothing highlights. That means the document is really a picture, a scan or a phone photo saved as a PDF, with no words underneath. You cannot search it, copy from it, or paste a quote into an email. OCR (optical character recognition) is the fix: it reads the shapes of the letters in the image and turns them back into real, selectable text. This guide shows how to OCR a scanned document and turn it into something you can actually use.

PDFMarkr's OCR PDF tool does this entirely in your browser. Your scan is processed on your own device and is never uploaded to a server, which matters when the file is a contract, a receipt, a medical form, or anything private.

Turn a Scanned PDF Into Text, Step by Step

Here is how to convert a scanned PDF to text and, if you want, keep the original look while adding a searchable layer:

1. Open the OCR PDF tool in any modern browser on Windows, Mac, Linux, iOS, or Android. No account, no install.
2. Drop in your file. It accepts scanned PDFs as well as images (.jpg, .png, .webp), so a photo of a page works too.
3. Decide whether to tick Also create a searchable PDF. Leave it on to get back a PDF that looks identical but has an invisible, selectable text layer added underneath.
4. Click Extract Text (OCR). On the first run only, the tool downloads a small English recognition model (about 11MB) and caches it for next time.
5. Watch the progress bar as it prepares each page and recognizes the text. Larger scans take longer because every page is read individually.
6. Review the extracted text, then copy it, download it as a .txt file, or download the searchable PDF.

That is the whole process. No sign-up, no watermark, and nothing leaves your machine.

Plain Text vs. a Searchable PDF

The tool can hand back two different results, and it helps to know which you want:

• Plain text (.txt) is best when you just need the words, to quote, reuse, or paste content somewhere else.
• A searchable PDF keeps the page exactly as it looks and layers the recognized text invisibly behind the image. The file still looks like the original scan, but now Ctrl+F works and you can select and copy from it. This is what most people mean when they want to make a PDF searchable.

Tips for the Best OCR Results

• Feed it the sharpest scan you have. OCR reads pixels, so a clean, high-contrast, straight scan at a decent resolution is recognized far more accurately than a dim, tilted phone photo.
• Straighten crooked pages first. If a page is rotated or sideways, fix it with Rotate PDF before running OCR so the lines sit level.
• Always proofread. No OCR is perfect. Skim the output for mangled numbers, split words, or stray characters, especially in totals, dates, and names.
• Check for an existing text layer first. If you can already highlight a word in the PDF, it is not a scan. Use Extract Text instead, which reads the real text directly and is more accurate than OCR.

Honest Limitations to Expect

Being straight about what OCR can and cannot do saves you frustration later:

• English only. The OCR PDF tool uses an English language model. Documents in other languages, or pages full of accented characters, will not be recognized reliably.
• Quality is capped by the scan. Blurry, skewed, low-resolution, or handwritten pages produce more errors, sometimes many. OCR estimates the most likely letters; it does not truly read the page.
• Layout may not survive. Complex tables, multiple columns, and sidebars come through as text, but the reading order can need manual cleanup afterward.
• It is an aid, not a certified transcript. Wherever accuracy is critical, verify the output against the original document.

Get Searchable Text Out of Your Scan

A scanned PDF does not have to be a dead end. Run it through the OCR PDF tool, choose plain text or a searchable PDF, and get back something you can search, copy, and reuse, free, private, and without uploading a thing. For more step-by-step walkthroughs across the toolkit, browse the full guides library.