Listen to any document.
Open a PDF, book or file and hear it read aloud with natural voices, following along sentence by sentence.
- Free, no signup
- Files never leave your device
- PDF, EPUB, Word, Markdown, text
Picking up where you left off.
How it works
- Private by design
- Your document is opened and read entirely inside this browser tab. Nothing is uploaded, and there is no account to make.
- Natural voices
- Speech comes from Gradium's open Phonon model running on your device. It downloads once (65 MB), then loads from the browser cache.
- Clean text
- Page numbers, running headers and end-of-line hyphens are removed from PDFs. EPUB books are read chapter by chapter, skipping the table of contents.
- Picks up where you stopped
- Your place in each document is remembered in this browser. Use Space to play or pause and the arrow keys to move by sentence.
Questions
Is DocRead free?
Yes. There is no account, no trial and no limit on document length.
Do I need to sign up?
No. Open a file and press play.
Are my files uploaded anywhere?
No. Your document is opened and read inside your browser, and the speech model runs on your own device. The only downloads are the site itself and the voice model, which comes from Hugging Face.
Which file types can it read?
PDF files with selectable text, EPUB books without DRM, Word documents (.docx), Markdown (.md) and plain text (.txt). You can also paste or type text.
Which languages and voices are there?
English only for now, with 37 voices: American, British and a set of character voices.
Can I download the audio?
Yes. Open the ⋯ menu in the player and choose Save audio to get the whole document as a WAV file.
Why does the first play take a moment?
The first time, your browser downloads the 65 MB voice model and keeps it in its cache, so later visits load in seconds. Each new sentence also takes a moment to generate before it plays; after that, DocRead keeps generating ahead of where you are listening.
Does it work on phones?
Yes, in current versions of Safari on iPhone and iPad and Chrome on Android. Phones generate speech more slowly than computers, so the first sentence takes a little longer.
Does it work offline?
Yes, after your first visit. Once you have played something, the site and the voice model are saved in your browser, so DocRead opens and reads without a connection. You can also install it: use Add to Home Screen on a phone, or the install button in the address bar of Chrome or Edge.
How much memory does it use?
DocRead keeps recently generated speech in memory so skipping back is instant: up to about 30 minutes of speech on computers (about 96 MB) and about 10 minutes on phones (about 32 MB), since phones reload tabs that use too much. Older speech is dropped and simply generated again if you go back to it. There is no limit on document length. Save audio is the exception: it holds the whole recording in memory while building the file, about 170 MB per hour of audio, so very long documents may be too large to save on a phone.
Does it remember where I stopped?
Yes. Your place in each document is saved in this browser, so opening the same file again picks up where you left off. Nothing is saved on a server.
Can it read scanned PDFs?
No. A scanned PDF is a set of page images with no text in it. Run it through OCR first (many scanner apps and PDF editors can), then open the result.