OCR PDF
Recognise text in scanned or image-based PDFs with in-browser OCR and add an invisible, selectable text layer over each page. Supports English, Nepali and Hindi.
Drag & drop your PDF
or choose a file below
PDF · up to 50 MB
How OCR works here
Upload a scanned PDF and choose a language. Each page is recognised with in-browser OCR, and an invisible, selectable text layer is added over the original image.
Languages and privacy
English, Nepali and Hindi are supported. OCR runs locally in your browser, so your document never leaves your device.
Step-by-step: making a scanned PDF searchable
- Drop the scanned PDF onto the upload area or choose it from your device. Files up to 50 MB are accepted.
- Pick the document language: English, Nepali, Hindi, or a combined option such as English + Nepali when a page mixes both scripts.
- Select “Make PDF searchable”. Every page is first rendered as an image and then recognised one at a time, and the progress bar shows which page is being read.
- When recognition finishes, a summary shows the number of pages, the words found and the size of the new file.
- Download the searchable PDF, saved with “-ocr” added to the original file name, and optionally the recognised text as a .txt file.
What the downloaded files contain
Each page of the searchable PDF is the original page rendered as a high-resolution JPEG image, with the recognised words placed on top as invisible text at the positions where they were found. The page looks the same as before, but you can now search it with Ctrl+F, select and copy passages, and let document management systems index the content. The .txt download holds the plain recognised text of all pages, separated by blank lines, which is useful when you only need the words for editing or pasting elsewhere. Because pages are rebuilt from images, existing digital text, links, bookmarks and form fields are not carried over, so use this tool on scanned or photographed documents rather than PDFs that already contain selectable text.
Choosing the document language
The language setting tells the recognition engine which alphabet and vocabulary to expect. Choose English for Latin-script documents, Nepali or Hindi for Devanagari text, and a combined option when one page contains both, such as a form with Nepali content and English headings. Combined options take longer and can be slightly less accurate on single-language pages, so use them only when needed. Language data is downloaded the first time you use a language, so the first page may take longer; the document itself is still processed on your device.
Nepali and Hindi documents
The invisible text layer uses a standard PDF font that covers Latin characters only, so Devanagari words are recognised but cannot be embedded in the searchable PDF. For Nepali or Hindi documents, use “Download text (.txt)” to get the recognised text in full. On mixed English + Nepali or English + Hindi pages, the English words are still searchable inside the PDF.
Tips for better accuracy
- Scan at around 300 dpi, or take sharp, evenly lit phone photos; blurry, skewed or low-contrast pages produce more mistakes.
- Fix sideways or upside-down pages with Rotate PDF before running OCR.
- Handwriting, decorative fonts, complex tables and very small print are recognised less reliably, so check names, dates and numbers in the result.
- Long documents take more time because every page is processed in your browser; split very large files with Split PDF and run them in parts.
Common uses
- Turning scanned contracts, receipts and letters into files you can search by name, date or amount.
- Making photocopied notes, old books and paper exam questions easy to search.
- Preparing scanned documents for archives or document management systems that need a text layer.
- Copying passages from a scanned report without retyping them.