Technology
Open-source Chrome extension OCR It extracts text from un-copyable documents for language models
The tool processes images locally using a bundled Tesseract engine, avoiding external network calls or API dependencies.
The short version
- A new open-source Chrome extension named OCR It enables text extraction from paginated, un-copyable web documents like scanned PDFs and slide decks.
- The extension runs locally using Tesseract OCR without making external network requests or sending data off the user's device.
- It features an automated loop that can capture screens, turn pages across complex web frames, and stop when page limits or duplicate content are detected.
Key facts
- OCR It is a Chrome extension designed to extract text from un-copyable documents and format it for language models.[Hacker News]
- The extension processes optical character recognition locally via a bundled Tesseract build without making outbound network requests.[Hacker News]
- It supports automated page captures that navigate documents across cross-origin iframes and shadow DOMs using coordinate points instead of CSS selectors.[Hacker News]
- Automated capture runs automatically terminate under specific conditions, including after two consecutive identical pages or reaching a 300-page limit.[Hacker News]