Paste something you would rather listen to than read — an article, a draft you are proof-listening, a passage in a language you are learning — and the browser speaks it using the voices already installed on your device.
A PDF or a Word document can be loaded straight in. A PDF works only if its text can be selected. A scan or a photograph of a page holds a picture rather than characters, and this tool does not run OCR, so there is nothing in such a file for it to read. Open the PDF and try to select a sentence: if the text highlights, it will be read.
Voices are labelled on device or network. An on-device voice reads the text without it leaving your machine. A network voice is generated by your browser vendor, which means the text is sent to them, so the tool says so before you press play rather than after.
How it works
- Paste text, or load a PDF, a .docx, or a text file.
- Choose a voice — on-device voices are listed first — and set the speed.
- Press read aloud, and pause or stop at any time.
Frequently asked questions
Which PDFs can be read?
Any PDF that carries a text layer, which is what you get from exporting rather than scanning. Open it and try to select a sentence: text that highlights will be read, and a page that behaves like one flat image will not. Utilark does not run OCR, so a scan is reported rather than guessed at.
Will it read the page numbers and the header?
No. A running header or footer that repeats across the document is recognised by that repetition and dropped, along with page numbers, because hearing the chapter title between two sentences is worse than not hearing it at all. A two-column layout is also detected, so each column is read to its end rather than the two being read across.
Is the PDF uploaded?
No. It is parsed in your browser. The one thing fetched from Utilark is a character-mapping table, and only when a PDF uses an older Korean or Japanese text encoding — that file is part of this site, and your document is not sent with the request.
Does my text get sent anywhere?
Only if you choose a voice marked network, which is generated by your browser vendor rather than your device. On-device voices speak without sending anything, and they are the default here.
Why does the voice list look different on my phone?
The list comes from the voices installed on your operating system, not from Utilark. Adding a voice in your system settings makes it appear here.
Why does long text play as separate pieces?
Browsers cut off a single long passage partway through, so the text is split at sentence boundaries and played as a queue. It also means stopping takes effect immediately.
Does it keep reading if I switch tabs?
Yes, so you can listen while doing something else. Leaving or closing the page stops it, because audio you can hear but cannot find is the one thing you cannot easily undo.