OCR PDF

Make scanned PDF pages searchable with OCR

Recognize printed text in scanned PDF pages directly in your browser, create a searchable PDF with a hidden text layer, and optionally download the recognized text as a TXT file.

Browser-based OCR

OCR PDF

Select one PDF, choose the OCR language and pages, then create a searchable copy.

PDF stays on your device

Drop one scanned PDF here or Choose PDF file Password-protected PDFs must be unlocked first.

Pages to process

All pages Check every page and apply OCR according to the selected OCR mode. Selected pages Enter page numbers or ranges such as 1-5,8.

OCR settings

OCR language The language model is downloaded to your browser when OCR starts. OCR quality Higher settings render larger page images and use more memory. PDF image quality Applies only to pages rebuilt for OCR.

OCR behavior

OCR scanned pages only Pages that already contain enough extractable text are copied unchanged. Image-only pages receive OCR. Recommended Force OCR selected pages Re-render and recognize every selected page, even when the original page already contains text.

Searchable PDF: OCR pages are rebuilt as page images with a text layer positioned behind the image. This lets compatible PDF readers search and extract recognized text while keeping the text layer visually hidden.

Why Smart mode is recommended: pages that already contain searchable text are copied directly from the original PDF instead of being rasterized again.

OCR is not perfect. Accuracy depends on scan resolution, language, font, contrast, page rotation, handwriting, tables, and image quality. Always review important names, numbers, dates, and legal or financial information against the original scan.

Rebuilt OCR pages: pages that require OCR are rendered to images and rebuilt with a hidden recognized-text layer. On those pages, original hyperlinks, forms, annotations, vector objects, layers, and other interactive PDF features may not be preserved.

Current searchable-text languages: this version is optimized for the Latin-script languages listed above. Other languages are not offered yet because producing a reliable searchable text layer requires appropriate font and script handling.

Privacy note: the selected PDF and rendered page images are processed in your browser. 2PDF2 does not intentionally upload them to its own server for OCR. The browser downloads the OCR engine, WebAssembly core, and selected language data from third-party code-delivery services. The webpage may also connect to analytics or advertising services as described in our Privacy Policy.

How to OCR a scanned PDF

Select Choose PDF file and open the scanned PDF.
Choose all pages or enter a page range.
Select the language used in the document.
Leave OCR scanned pages only selected unless you specifically need to force OCR on existing searchable pages.
Select Run OCR and keep the browser tab open while recognition runs.
Download the searchable PDF and, when useful, the extracted TXT file.

What is OCR?

Optical Character Recognition analyzes the pixels in a scanned page and attempts to identify printed letters, words, and lines. A normal image-only scan may look like text but behaves like a photograph. OCR adds machine-readable text so compatible PDF readers can search, copy, and extract recognized words.

How 2PDF2 creates a searchable PDF

Detect existing text

In Smart mode, the tool first checks selected pages for existing extractable text. Pages that are already searchable are copied directly into the output PDF.

Render scanned pages

Image-only pages are rendered at an OCR-friendly resolution using PDF.js before recognition begins.

Recognize words

Tesseract.js analyzes the rendered page and returns recognized text and word-position information.

Add hidden text layer

The recognized words are positioned in the PDF and then covered by the original page image, creating a visually unchanged image page with machine-readable text underneath.

Which OCR quality should I use?

Standard is suitable for most clear scans. High accuracy renders a larger page image before recognition and can improve results on small print, but it needs more memory and processing time. Fast mode is useful for large documents when speed matters more than maximum recognition quality.

Why can OCR make a PDF larger?

Image-only pages that require OCR are rendered and rebuilt as compressed page images with an additional text layer. Depending on the source PDF and selected quality, the resulting searchable PDF may be larger or smaller than the original.

Does OCR preserve the original PDF exactly?

Pages that Smart mode detects as already searchable are copied directly. Pages that need OCR are rebuilt visually from rendered page images. This means interactive features on rebuilt pages can be lost even though the visible page is retained.

Frequently asked questions

Is the OCR PDF tool free to use?

Yes. The current 2PDF2 OCR PDF tool can be used without a paid account.

Is my PDF uploaded to 2PDF2?

The PDF and rendered page images are processed in your browser and are not intentionally uploaded to a 2PDF2 OCR server. OCR libraries and language models are downloaded from external code-delivery services.

Can OCR make scanned text searchable?

Yes. Recognized words are added to rebuilt scan pages as a hidden text layer so compatible PDF readers can search and extract them.

Does OCR work on handwriting?

Tesseract-based OCR is primarily intended for printed text. Handwriting accuracy can be poor and should not be relied on.

Can I OCR only a few pages?

Yes. Choose Selected pages and enter ranges such as 2-6,10. Pages outside the range are copied without OCR.

Why was a page skipped in Smart mode?

Smart mode skips OCR when the page already contains enough extractable text. Use Force OCR selected pages when you specifically want the page rendered and recognized again.

Can I download the OCR text separately?

Yes. After OCR finishes, the tool provides a TXT download containing the extracted or recognized text for the selected pages.

Can I OCR a password-protected PDF?

Not directly. Unlock an authorized protected PDF first, then run OCR on the unencrypted copy.

Built for transparent browser OCR

2PDF2 distinguishes between pages that already contain searchable text and pages that actually need OCR, reports recognition confidence, provides extracted text for review, and clearly explains which pages are rebuilt. For questions, bug reports, or feedback, visit our Contact page.