Aloud: Natural Text to Speech Reader, Offline

Reads any article or PDF aloud in a natural neural voice that runs on your device. Offline and private, it highlights every word.

Listen to the web in a natural voice that lives on your device. Aloud reads any article or PDF aloud and lights up each sentence, and with most voices each word, as it's read. The voice is a neural network running inside your browser, so it works offline and nothing you read is sent anywhere, unless you choose to use a speech server of your own. No account, no subscription, free and open source. • 324 natural voices in 52 languages, from four neural engines: Phonon, Supertonic 3, Kokoro-82M and Piper • Reads the article, not the menus and ads, and highlights the sentence and the word in place; documentation too, with its notes, tips and code • A page in two languages? Each passage is read by a voice that speaks it • Reads pages that show their text in an embedded page (Claude artifacts) or inside web components (MSN) • The whole page, a selection, or from any sentence: with a click, a right-click or a key • PDFs too: real pages, highlighted where the text sits; mark them up, kept on your device, and save the marks into the file • Picks up at the sentence you stopped at; a reading queue for later; save any article as an MP3 • Highlights, notes and tags on any website, kept on your device and exportable as Markdown • The same keys for reading and for any video or audio on the web: S slower, D faster, R normal speed, Z back, X forward, M mute. Alt+M turns the video control on or off for a site, and it can stop videos that play by themselves • Every key can be changed. Light and dark themes. • Sync through your own Dropbox: your notes, highlights, settings and the place you stopped reading follow you to every device, phones included. Off until you connect it, and it goes to your Dropbox and nowhere else. Private by construction: there is no Aloud server. Voices download once from Hugging Face; after that, reading needs no connection at all. Bugs and ideas: https://aloud.souj.dev/feedback.html Source code: https://github.com/smailoujaoura/aloud