OCR PDF

Extract plain text from scanned PDFs using optical character recognition in your browser.

Smart OCR

Pages with existing text are detected and extracted directly. Only scanned pages undergo OCR, saving time and improving accuracy.

Drag & drop your PDFs here

or click to browse your files

Max 50 MB per file

Document Language

First use may take longer as OCR models are downloaded (~15MB).

Capabilities and limits

Accepts: PDF
Produces: Plain text
Limits: 50 MB per file; Eight selectable OCR languages; Does not create a searchable PDF

Reproduce this OCR PDFs example

Run English OCR on the image-only fixture.

Reviewed fixtures

1-page synthetic PDF — 4059 bytes

Expected facts

The result contains the reviewed ground-truth phrase “Test Image”.
This is one fixture result, not a general OCR accuracy benchmark.

Proof owner: tests/tool-workflows.spec.ts — recognizes the known scanned-fixture phrase and observes network requests

Open OCR PDFs Verify fixture hashes and method

What happens to your document

1. Browser inputYour browser reads the file you select. Document bytes are not sent to a LocalPDF processing API.
2. Local resultThe tool performs the operation in this tab and creates its result in browser memory.
3. Your downloadYou choose whether to download the result. Browser memory is cleared when the page is closed or reloaded.

Why Process PDFs Locally?

Supported workflows process document bytes in the browser without a document-byte upload to LocalPDF. Device, browser, extension, download, storage, and later-sharing risks remain separate boundaries.

Review the no-upload verification method →