OCR PDF online
Make scanned PDF text searchable and selectable with OCR that runs privately in your browser.
Drop a scanned PDF here, or click to choose one
Your PDF is processed locally. Clear, upright English scans work best.
OCR PDF
Ready for local OCR
Choose your settings, then select Run OCR.
- Selected pages get searchable text.
- Unselected pages remain in the document.
- Your file is not uploaded.
Searchable PDF is ready
Page summary
Search scanned pages
Add a hidden text layer so words in image-only pages can be found and selected in a PDF reader.
Keep useful text
Smart OCR leaves pages with existing selectable text unchanged and recognizes the pages that need it.
Process locally
PDF.js, Tesseract, and PDF creation code work in your current browser tab without sending the document away.
How to OCR a PDF online
- 1
Add a scanned PDF
Drop one PDF into the upload area or choose a document from your device.
- 2
Choose OCR settings
Select Smart OCR or Force OCR, choose a page range, and select balanced or high quality.
- 3
Run OCR in your browser
The selected pages are rendered, recognized with the local English model, and rebuilt with searchable text.
- 4
Download the searchable PDF
Review the page summary, save the new PDF, and test important words in your PDF reader.
What makes a searchable PDF different
A scanned PDF often contains one photograph of paper on each page. It can look perfectly readable to a person while a computer sees no letters at all. OCR analyzes the pixels, estimates the printed characters, and places an invisible text layer over the page image. A compatible PDF reader can then search, select, and copy that recognized text while showing the original page appearance.
OCR does not turn the page into an editable word-processing document. The visible page remains an image, and the recognized text is aligned behind it for search and selection. Use Extract text from PDF when you need a separate TXT result that you can review and edit directly.
Smart OCR and Force OCR
Smart OCR
This is the recommended choice for mixed documents. The tool checks each selected page and preserves pages that already contain useful selectable text. Image-only pages receive OCR. Preserved pages keep their vectors, links, form fields, annotations, and existing text quality.
Force OCR
This mode recognizes every selected page. It can help when an old or incorrect hidden text layer makes search unreliable. Because processed pages are flattened, use it only when replacing the existing page text is more important than keeping interactive page elements.
Balanced and high quality recognition
Balanced mode renders each selected page at 144 DPI. It is suitable for ordinary letters, forms, reports, and scans with medium or large print. High quality mode uses 216 DPI so the OCR engine receives more detail around small characters. It requires more memory, takes longer, and can increase the size of the searchable PDF.
On a phone or tablet, start with Balanced mode and a short page range. Browser memory is limited, especially when a source page is unusually large. The tool releases each page image after recognition, but one high resolution page and the OCR engine must still fit in memory at the same time.
Private OCR in your browser
The source PDF is opened from the file you selected in this tab. PDF.js renders selected pages to temporary canvases, the local Tesseract WebAssembly worker recognizes the English text, and PDF creation code assembles the downloadable result. The page canvases are discarded as processing continues.
The first OCR run may download the engine, its WebAssembly core, and the English language model from LocalPDFLab. Those static resources can be cached by your browser. The resource requests do not contain your document, filename, page images, or recognized text.
How to improve OCR accuracy
- Use sharp scans with strong contrast between the text and background.
- Rotate sideways or upside-down pages before starting OCR.
- Choose High quality when small print is unclear in Balanced mode.
- Avoid shadows, folds, heavy compression, and blurred camera images.
- Use Smart OCR when some pages already have accurate selectable text.
- Verify names, account numbers, totals, dates, and legal wording manually.
Limits of browser based PDF OCR
Recognition quality depends on the source. Handwriting, decorative fonts, mathematical notation, dense tables, multi-column layouts, curved pages, and faint photocopies can produce mistakes. The confidence score is a useful signal, but it does not prove that every word is correct.
Processed pages are rebuilt as searchable page images. Interactive form fields, clickable links, comments, layers, embedded media, and some accessibility structure on those pages may not remain interactive. Pages skipped by Smart OCR and pages outside your selected range are copied from the source PDF instead.
Frequently asked questions
What does OCR do to a PDF?
OCR examines the image on each selected page, recognizes printed characters, and creates a searchable text layer. The visible scan remains on the page while supported PDF readers can search, select, and copy the recognized words.
Is my PDF uploaded to a server?
No. PDF rendering, character recognition, and output creation happen in your browser. The OCR engine and English language model are static files downloaded from LocalPDFLab, but your PDF and recognized text are not sent with those requests.
Can I OCR only certain pages?
Yes. Leave the page field blank to process the full document, or enter pages and ranges such as 1-5, 8. Pages outside the chosen range are copied into the result without OCR.
What is Smart OCR mode?
Smart OCR checks selected pages for useful selectable text. Pages that already contain text are kept unchanged, while image-only pages are recognized. Force OCR processes every selected page, even if a hidden text layer already exists.
Which languages are supported?
This version uses an English printed text model. Documents in other languages may produce incomplete or incorrect text. More language models can be added later without changing the local processing approach.
Does OCR work on handwriting?
The engine is intended for printed text. Neat block handwriting can sometimes produce partial results, but handwriting recognition is not reliable enough for important documents.
Why can the searchable PDF be larger?
A processed page contains a rendered page image plus a hidden text layer. High quality mode renders more pixels for small characters, which can improve recognition but uses more memory and can increase the output size.
Will forms and links still work?
Pages processed with OCR are flattened into searchable page images, so interactive form fields, links, annotations, and layers on those pages may no longer be interactive. Smart OCR keeps pages with existing text unchanged, and unselected pages are also preserved.
Next steps
Related PDF tools
Choose another private, browser-based tool for the next step in your document workflow.