lucid — Private Document Reader BETA

THIS EXTENSION IS IN OPEN BETA. lucid is a private document reader that turns PDFs, scanned textbooks, and text documents into a guided listening experience. Open a paper, report, textbook, Markdown file, transcript, or saved article and follow along while a natural neural voice reads it aloud. Document processing, printed-text recognition, and speech generation happen locally on your device. lucid does not upload your documents or extracted text to a lucid server. Features include: • Private, on-device neural speech • 28 American and British English voices • Original-page and clean-text reading views • Live paragraph, sentence, and estimated word tracking • Local OCR for printed scanned textbooks • Support for conventional multi-column documents • Chapter and reading-order detection • Suppression of common page numbers, headers, figures, and tables • Highlights, comments, and document notes • Reading-position and recent-document resume • Adjustable playback speed • PDF, Markdown, plain-text, and saved HTML support • Per-document private mode • Local data and voice-cache controls Scanned books can be recognised in selectable page ranges. Additional pages may be processed in the background while you continue reading. Privacy is a core part of lucid. PDFs, recognised text, annotations, notes, and generated speech remain on your device. Remote PDFs are fetched directly from their original website. A neural voice model is downloaded when first needed, but document content is not included in that request. Current beta limitations: • OCR is intended for clear, upright printed English—not handwriting. • Complex layouts, decorative text, and heavily damaged scans remain best effort. • Neural speech requires a one-time model download. • Highlights and notes are stored locally in lucid and are not written into the original document. The open beta is currently free, includes all available features, and does not require a lucid account.