OCR It – pull text out of un-copyable documents for your LLM (github.com)

A new Chrome extension, OCR It, has been launched to revolutionize the way users extract text from un-copyable documents, such as scanned books and locked PDFs. By allowing users to define a capture region just once, the extension can automate the process of taking screenshots, running Optical Character Recognition (OCR) using a local Tesseract build, and compiling the extracted text into a single transcript with simple hotkeys. Users can also initiate an auto-run mode, enabling the extension to seamlessly turn pages and continue capturing text without any manual intervention, which significantly enhances productivity for those needing to convert extensive text materials for analysis or use with large language models (LLMs) like Claude or ChatGPT. This extension is significant for the AI/ML community as it democratizes access to valuable textual data embedded in otherwise inaccessible formats. By providing local OCR capability without the need for API keys or internet connectivity, it ensures user privacy and data security, as all processing remains on the user's machine. The technical implications include support for multiple languages and efficient page handling via hotkeys, along with a sophisticated mechanism for dealing with cross-origin iframes and shadow DOMs—common challenges in web scraping. This makes OCR It an invaluable tool for researchers, educators, and anyone dealing with large volumes of textual content that require transformation into usable digital formats.