OCR PDF

Turn a picture of text into text you can actually use

A scanned PDF is just a stack of photographs. Your eyes read it fine, but to the computer each page is a flat image with no characters inside it, so Ctrl+F finds nothing and you can't highlight a sentence. OCR (optical character recognition) reads the shapes in those images, recognizes the letters and words, and writes them back as a real text layer positioned exactly behind the picture. The page still looks identical; it just stops being "dumb pixels."

Where this actually matters

• You inherited a folder of scanned contracts or invoices and need to find every mention of a vendor or clause number. After OCR, one search across the document returns hits instead of you scrolling page by page.
• A research or legal team wants to quote from a scanned report. Run OCR, then select and copy the exact paragraph rather than retyping it.
• Accessibility / compliance. Screen readers, and most "is this PDF accessible?" checks, require a text layer. A purely image-based scan fails; an OCR'd one passes the first hurdle.
• You're feeding documents into other software — contract management, e-discovery, or full-text indexing — that can only ingest selectable text. OCR is the prep step that makes the batch usable.

What's happening under the hood

This tool runs Tesseract OCR (via the ocrmypdf pipeline). It does not flatten or redraw your page — it adds an invisible text layer aligned to the original image, so the document looks pixel-for-pixel the same while becoming searchable. English is the recognition language by default, and the engine works best when that language matches the document's actual script.

Practical tips for better results

• Resolution is everything. Roughly 300 DPI scans give Tesseract clean letter shapes; a 72-DPI screenshot or a heavily compressed scan will produce errors no engine can rescue.
• Straighten and de-noise first. Skewed, shadowed, or speckled scans hurt accuracy. A crisp, high-contrast black-on-white page is ideal.
• OCR adds text, it doesn't fix layout. Multi-column pages and dense tables are recognized but won't necessarily land in reading order.

Honest limitations — and when to reach for a different tool

OCR makes a scan searchable and copyable. It does not make the scanned image editable — you cannot click into a word and rewrite it, because the picture underneath is unchanged.

• Need an editable, reflowable document for Word? Run [PDF to Word](/pdf-to-word) after OCR — for scanned files, OCR first is what gives the converter real text to extract.
• Want a long-term archival format that locks in fonts and structure for records retention? Use [PDF to PDF/A](/pdf-to-pdfa).
• Only need a handful of pages searchable? [Extract PDF pages](/extract-pdf) first, then OCR the smaller file.
• Want to mark up or hide parts of a scan? That's [Edit PDF](/edit-pdf) — it stamps an overlay and never touches the underlying text layer.

If your PDF already has selectable text (it was exported from Word or generated digitally), you don't need OCR at all — searching already works.