OCR PDF: Make Scanned PDFs Searchable

A scanned document is just a picture of a page: you cannot search it, select a sentence or copy a figure out of it. PDF OCR (optical character recognition) reads the letters in that picture and turns them into real text. You can keep the original look and add an invisible text layer, which gives you a searchable PDF, or simply download the recognized words as a .txt file. Recognition runs in your browser with the open-source Tesseract engine, so your scan never leaves your device.

How to OCR a PDF

1. Add a scanned PDF or an image (PDF, JPG, PNG, WebP or BMP, up to 100 MB). You can also paste a screenshot with Ctrl+V.
2. Choose the document language. It defaults to the language of this page, and you can add a second language for mixed documents, for example English and French.
3. Pick the output: Searchable PDF or Plain text. For a PDF you can limit OCR to specific pages, such as 1-3, 7.
4. Click Recognize text. The first run downloads the language data, which is then cached for next time. Progress is shown page by page.
5. Download the result: yourfile-ocr.pdf or yourfile.txt.

Supported languages

English, French, Spanish, German, Italian, Portuguese, Russian, Arabic, Hindi, Chinese (Simplified), Chinese (Traditional), Japanese, Korean, Turkish and Urdu. Selecting the right language matters: it tells the engine which alphabet and which words to expect, and a wrong choice produces nonsense. Only pick a second language when the document really contains both, because every extra language makes recognition slower.

What the searchable PDF looks like

Your pages are not changed or re-compressed. Each page is rendered at up to 300 DPI for recognition, and the recognized words are then placed as an invisible layer exactly over the printed ones. The document looks the same as before, but Ctrl+F finds words, you can highlight and copy passages, and search tools and document management systems can index the file. The plain-text option puts the text of each page under a small page marker, which is handy for pasting into an email or a spreadsheet.

How accurate is it?

• Best results: clean, printed text scanned at 300 DPI, straight pages and good contrast. Here Tesseract usually gets nearly every word right.
• Weaker results: blurry phone photos, very small print, tilted or curved pages, colored backgrounds and decorative fonts. Rescan or crop if you can.
• Not suitable: handwriting. Tesseract is trained on printed type and does not read handwritten notes reliably.
• Complex scripts such as Arabic, Hindi and Chinese generally need a sharper scan than Latin text to reach the same accuracy.

Always proofread important numbers, names and dates, as with any OCR.

Tips and related tools

• Photographing paper documents? Scan to PDF straightens and cleans up pages before you OCR them.
• Want structured text instead of a flat .txt? OCR the file first, then convert the searchable PDF with PDF to Markdown.
• If the PDF already contains selectable text, OCR is not needed: the PDF text extractor is instant and exact.
• Password-protected files must be unlocked first with Remove PDF Password.

Private and free

Because recognition happens on your own computer or phone, confidential scans such as contracts, IDs and medical letters are never uploaded to WoowPDF. Only the language model is downloaded. There is no sign-up, no page limit per day and no watermark. Speed depends on your device: expect a few seconds per page on a laptop, longer on older phones.