Extract text from a PDF or an image (OCR)

Text recognition inside your own browser: neither the document nor your photos leave your device.

A scanned document, a PDF whose text has been converted to vector strokes, or an image (JPG, PNG or a photograph) contain no characters, only pixels: perfectly legible to the eye, but not to a program looking for text.

Those cases need OCR, optical character recognition. Here it runs on Tesseract compiled to WebAssembly, executing inside your browser, and pages are rasterised with the same PDF library that ships in Firefox. Free, no sign-up: there is no third-party computer vision service involved, and no per-page billing.

This matters because the documents typically put through OCR are payslips, contracts, invoices, medical reports or scanned identity documents. Sending them to an online converter means placing the entire document on a third-party server and, where it contains personal data, constitutes processing that must be justifiable. Here no such processing takes place: the file never leaves the device.

OCR is never perfect: with clean printing and straight pages it gets most of it right, but handwriting, skewed photos or degraded scans are worth reviewing.