OCR PDF
Updated Files never leave your device
Scanned PDFs are just pictures of pages, so you can't search, select or copy their text. OCR (optical character recognition) reads the letters and gives you both a searchable PDF and the plain text — all on your device.
How to OCR a PDF
- Choose a scanned PDF.
- Select the document's language(s). The model downloads once (1–4 MB) and is cached.
- Optionally limit the pages (e.g. 1-5) and pick Fast or Accurate.
- Click Run OCR and watch each page's progress — Cancel at any time.
- Download the searchable PDF, and copy or download the recognized text.
What you get
- A searchable PDF that looks exactly like the original — the scanned page images are untouched — with an invisible text layer placed over each word, so Ctrl+F, copy-paste and screen readers work.
- Plain text of every page, separated by page markers, ready to paste into a document or download as .txt.
- A confidence score per page, so you know which pages to proofread.
Which scan quality to choose
| Setting | Best for | Speed |
|---|---|---|
| Fast (200 DPI) | Clean office scans, 11 pt text or larger | ≈2× faster |
| Accurate (300 DPI) | Small print, footnotes, faxes, phone scans | Recommended default |
Tesseract reads best when letters are 20–30 pixels tall. A 10 pt font is about 42 px tall at 300 DPI (10 ÷ 72 × 300) but only 28 px at 200 DPI — which is why small print needs the Accurate setting.
Mixed PDFs
Many PDFs contain both digital pages (with real text) and scanned pages. With “Skip pages that already have text” on, digital pages keep their original, perfect text and are not OCR'd again, which is faster and avoids duplicate text when you search.
Private by design
PDF Fox renders your pages with pdf.js and recognizes text with Tesseract, both running in your browser. Your document is never uploaded; the only download is the open-source language model for the languages you select, which your browser caches.
Related tools
Frequently asked questions
Is my PDF uploaded?
No. Pages are rendered and recognized inside your browser (pdf.js + Tesseract in WebAssembly). The only download is the OCR language model, once per language; your document never leaves your device.
How long does it take?
Roughly 3–10 seconds per page on a laptop at 300 DPI, longer on phones. Use the page range to OCR only the pages you need, or Fast (200 DPI) for clean, large print.
Will the PDF look different after OCR?
No. The original pages are kept as they are; the text layer is invisible (PDF text rendering mode 3). File size grows only slightly — usually a few KB per page.
Can it OCR handwriting?
No. Tesseract is trained on printed text, so handwritten notes and signatures won't be recognized reliably.
Is there a page limit?
No fixed limit — it's bounded only by your device's memory and patience. For long documents, OCR in batches with the page range.
Your business here? Sponsor this tool from $9/week. See placements