AI Signal 142
Chrome extension OCRs paginated documents locally, outputs text for LLMs
OCR It is a Chrome extension that captures and OCRs text from paginated documents trapped in viewers that prevent selection, running entirely locally with a bundled Tesseract build.
Engineers working with scanned books, slide decks, or PDFs in restrictive viewers can extract text without manual transcription or cloud OCR services. The local-only processing means sensitive documents never leave the machine, and the output is immediately usable by LLMs for summarization or search.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
OCR runs locally via bundled Tesseract with no outbound network requests, requiring no API key.
Users pin a capture region once and either manually capture each page with a hotkey or let the extension auto-advance through the entire document.
Auto-advance uses screen coordinates rather than CSS selectors to handle cross-origin iframes and shadow DOM, but cannot reach Chrome's built-in PDF viewer or plugin-based readers.
THE CLUSTER