How to extract text from a scanned PDF (OCR)

Scanned PDFs are really a series of images — photographs or scans of paper pages. You cannot copy, search, or edit the words until optical character recognition (OCR) converts those images into actual text. This guide shows you how to run OCR on a scanned PDF free in your browser, keeping the file on your device the entire time.

Scanned PDF vs. text-based PDF

A text-based PDF stores characters you can highlight and copy — for example, a report exported from Word or a web page saved as PDF. A scanned PDF stores pictures of pages. If you try to select text and nothing highlights, or copy-paste gives gibberish, you likely have an image-based document.

The PDF to Text tool extracts embedded text from normal PDFs quickly. For scans, you need OCR PDF, which renders each page as an image and runs Tesseract.js recognition locally. No upload servers are involved.

What OCR can and cannot do

OCR works best on clear, high-contrast scans with straight pages and readable fonts. It struggles with handwriting, heavy watermarks, very low resolution, skewed photos, and decorative typefaces. Results are rarely perfect — expect to proofread the output, especially for numbers, tables, and proper nouns.

OCR produces plain text (a .txt file), not a new searchable PDF with an invisible text layer. You can paste the text into Word, Google Docs, or a notes app. If you need images of individual pages instead, use PDF to JPG.

Step-by-step: OCR a scanned PDF online

  1. Open the OCR PDF tool. Go to OCR PDF in Chrome, Firefox, Safari, or Edge. Allow the page to load fully so the OCR engine is ready.
  2. Upload your scanned PDF. Drag the file onto the drop zone or click to browse. If the PDF is password-protected, unlock it first with Unlock PDF when you know the password.
  3. Select the document language. Choose the language that matches the printed text (for example, English). Correct language selection significantly improves accuracy.
  4. Click “Run OCR.” Each page is rendered and analyzed one at a time. Progress updates show which page is being processed. Large or high-resolution scans take longer because every pixel is examined.
  5. Review and download the text. When OCR finishes, preview the extracted text in the box on the page. Click download to save a .txt file. Edit and correct any misread words in your text editor.

Practical tips for better OCR accuracy

Start with a clean scan

If you control the source document, scan at 300 dpi or higher, keep pages flat, and avoid shadows from phone cameras. Crop excess margins with Crop PDF before OCR so the engine focuses on text.

Pick the right language

Mixed-language documents are harder. Run OCR once per language section if accuracy drops, or process the most common language first and manually fix foreign phrases.

Work in smaller chunks

Very long PDFs consume browser memory and time. Use Split PDF or Extract pages to OCR one chapter at a time on slower devices.

Reduce file size for faster processing

Oversized scan PDFs slow OCR down. Compress PDF can shrink page images while keeping text readable enough for recognition.

Proofread numbers and names

OCR often confuses 0 and O, or 1 and l. Double-check invoice amounts, dates, ID numbers, and email addresses before relying on the output.

Privacy and offline use

Because OCR runs in your browser, scanned contracts, medical forms, and personal records never leave your machine. You need an internet connection to load the tool the first time (JavaScript libraries are fetched from CDNs), but your PDF content is not sent for processing. After the page is cached, you may be able to OCR without an active connection on some browsers.

OCR PDF free — open the tool

Improve the scan before recognition

OCR cannot recover detail that the image never captured. Use sharp pages, sufficient resolution, high contrast and even lighting. Rotate and deskew sheets, remove dark scanner borders and select the correct language. Test a representative page containing small text, columns and numbers before processing a long document.

Review the errors OCR commonly makes

Similar characters such as O and 0 or l and 1, accents, punctuation and hyphenated line endings are frequent sources of errors. Tables may be read in the wrong order and handwriting may be unreliable. Compare names, dates, totals, account details, legal clauses and citations directly with the image; high-stakes text requires human verification.

Searchability is not complete accessibility

A recognized text layer can enable search and selection while the original page image remains visible. It does not automatically supply correct headings, reading order, table structure, document language or alternative descriptions. A document intended for assistive technology needs a separate accessibility review and, where necessary, remediation in an appropriate authoring or PDF accessibility tool.

How to test the result before relying on it

Do not judge the result only from the preview inside the tool. Download it, open it in a separate PDF reader and compare it with the source. Check the page count, first and last page, small text, rotated pages, links and any fields or signatures. If it will be printed, print one representative page at actual size. If it will be submitted to a portal, verify the portal’s accepted size, page format and security requirements before the deadline.

A safer workflow for important documents

Keep the original unchanged and give each output a descriptive name that includes its purpose or version. For work, legal, financial, medical or identity documents, follow the responsible organization’s rules for authorization, storage and retention. Local browser processing means the selected file is not uploaded to PDF Online Free for the operation, but downloaded copies, browser history and shared devices still need appropriate care.

Related tools

  • PDF to Text — fast extraction when the PDF already contains selectable text.
  • PDF to JPG — export pages as images for manual review or editing.
  • Compress PDF — reduce scan file size before running OCR.

Frequently asked questions

Does this work on scanned PDFs? Yes. The tool renders each PDF page as an image and uses OCR to recognize visible text.

Is my PDF uploaded for OCR? No. The PDF is processed locally in your browser and is not sent to a server.

Why does OCR take longer than normal text extraction? OCR analyzes the image content of every page, so large PDFs and high-resolution scans can take several minutes.