Hacker News
OCR It Pins a Screen Region and Hotkeys Through Paginated Viewers, Running Tesseract Fully Offline in the Browser
A 101-star MIT-licensed Chrome extension (Manifest V3, Chrome 116+) for pulling text out of documents that block selection, such as scanned books, locked PDFs and slide decks. It bundles a tesseract.js build supporting roughly 100 languages with English, Portuguese and Spanish preinstalled, and the author states OCR runs 100% offline with no API key, no network calls and no images leaving the machine. The workflow is the useful part for anyone feeding locked documents to an LLM: pin the region once, then advance pages by hotkey while it compiles a single transcript.
↳ Follow the thread