Run OCR on every page, output a PDF that looks the same but has a hidden searchable text layer.
Drop a scanned PDF here or click to browse
First run downloads language data (~10-30 MB) — subsequent OCR is faster.
Ready to start OCR
What "searchable PDF" means
A searchable PDF looks identical to the original scan but has invisible text positioned over each scanned word. Search tools (Ctrl-F in Acrobat / Preview / browsers), copy-paste, and screen readers can all access the text.
When this works well
Clean scans of printed text — 200+ DPI, straight pages, good contrast — accuracy 95%+
Modern documents — books, reports, invoices, modern letters
When it struggles
Handwritten content — Tesseract is print-trained; accuracy drops sharply
Low-resolution / skewed scans — pre-process first (straighten, sharpen)
Decorative or display fonts — Tesseract trained on standard text fonts
Tables and forms — text is extracted but layout / column structure is lost
Privacy
Tesseract.js runs entirely in your browser via WebAssembly. The OCR engine and language files are loaded from jsDelivr on first use. Your PDF and the OCR'd text never leave the page.