Rohman
You press Ctrl+F, type a word you can see on the page, and get zero results. The search is not broken. There is no text in the file, only a photograph of text, and a photograph has nothing to match against.
The Five-Second Test
Before reaching for any tool, find out which PDF you have. Try to select a word with the mouse. If the cursor grabs individual letters, the file has a real text layer and search should already work. If the whole page highlights as one block, or nothing selects at all, you are looking at a scan.
Two edge cases worth knowing. A PDF can be partially searchable (the cover page OCR'd, the rest images), so test a middle page, not the first. And a rare few files have text drawn as vector outlines: it looks selectable but copies as gibberish. If your file fails the test, the fix is OCR.
What "Searchable" Actually Means
OCR does not retype your document. It keeps the scanned image exactly as it is and adds an invisible text layer on top: each recognized word stored as invisible text positioned over the pixels it came from. When you highlight a "word" in a searchable scan, the highlight rides that hidden layer.
Acrobat's own output modes make the idea concrete. Searchable Image keeps the picture and adds the layer. Searchable Image (Exact) keeps the original image untouched, no resampling; the mode legal and forensic workflows prefer. Editable Text and Images replaces the picture with real rendered text, changes how the page looks, and is Pro-only.
The consequence that surprises people: after OCR, the PDF looks identical. The only change is that Ctrl+F, copy-paste, and screen readers now work. The file grows by a few tens of KB per page; the layer is small.
The Free Routes, and Their Real Limits
Google Docs will OCR a PDF for free: right-click the file in Drive, open with Google Docs. The official limits are a 2 MB file cap, text at least 10 pixels tall, and a warning that lists, tables, columns and footnotes are "not likely to be detected." Long files get silently truncated; community reports put the cutoff somewhere around 65–80 pages. And the output is a Google Doc, not a searchable PDF: you get the text, you lose the page.
PDF24 has a free web OCR tool and a free desktop app, both built on Tesseract. The web version uploads your file; the desktop one does not.
OCRmyPDF is the power-user answer: free, open source, wraps Tesseract and Ghostscript, can deskew and clean the scan, and can emit PDF/A. It is a command-line install, which is the catch.
The paid tier: Acrobat Standard at $14.99/mo covers the searchable modes, with the editable mode behind Pro at $19.99/mo. Smallpdf puts OCR behind Pro at about $12/mo. iLovePDF lists its OCR PDF tool on the free tier, with the OCR converters Premium at $5/mo billed annually. Every one of those is a server upload; the scan leaves your machine.
Why OCR Misses Words You Can See
OCR writes guesses into the layer, and the page still looks perfect because you are seeing the original image. The errors are invisible until they matter: rn read as m, l/I/1 swapped, a misread digit in an invoice total that nobody catches until someone searches for it.
Three things decide accuracy. Resolution: Tesseract's own documentation says it works best at 300 DPI or better, and below that, rescaling the image first helps. Skew: a tilted page breaks line segmentation badly enough that deskewing is worth doing before OCR, not after. Language: the engine guesses against a dictionary, so a German document run through the English pack produces confident nonsense.
Tables are the known weak spot: Tesseract's documentation admits the problem outright. Verify totals on an OCR'd invoice before trusting them.
Searchable Is Not PDF/A
The two terms get conflated because tools like OCRmyPDF can produce both at once. They are different things. Searchable is a property: the file has a text layer. PDF/A is ISO 19005, an archival standard: fonts embedded, no external references, no encryption, so the file renders identically decades from now.
The practical rule: if a court or an archive asks for PDF/A, searchable alone is not enough. If you just want Ctrl+F, you do not need PDF/A.
Making One Now
- Open OCR PDF
- Drop the scan in. It processes in your browser, so a medical record or a contract never leaves the machine
- Search a few words you can see on the page, and copy a paragraph out, to sanity-check the layer
If the scan is also too big to send once it is searchable, Compress PDF handles that, and this post explains which rung of the ladder to start on. For a scan that needs actual edits rather than just a text layer, Edit PDF is the next step.
The scan stays a picture. Searchable just means the picture finally has an index.
Make a PDF searchable — in your browser, nothing uploaded.
