Skip to content
CornerTools

Text to Speech

Listen to any text read aloud by your device's voices, with the sentence being read highlighted.

1×
1

Reads text aloud with the voices your browser offers, highlighting the sentence being read and, where the voice reports it, the word. Paste text or open a text, Markdown, PDF or Word file; files are read in your browser. Each voice is labelled with where it speaks: on your device, or online, where the text is sent to a speech service.

How to use it

  • Text and files — type or paste text, drop a file on the box, or use Open file. A Markdown file is read as the text it shows, without its markup; a Word file (.docx) as its paragraphs; a PDF as the text in it. A scanned PDF is pictures of pages with no text in it: run it through Searchable PDF (OCR) first. A password-protected PDF opens once Unlock PDF has removed its password.
  • Reading — Play shows the text sentence by sentence and reads it. Tap a sentence to read from there, or step with Previous and Next. Pause stops the voice; Resume reads on from the last word spoken, or from the start of the sentence where the voice reports no words. Edit goes back to the text box with the cursor where the reading was.
  • Voice — voices on your device come first, each with its language. Until you pick one, the voice follows the text: one for its script (a Hindi voice for Devanagari, where there is one), otherwise one for your browser's languages. A voice on your device is chosen before an online one.
  • Speed from 0.5× to 2× and Pitch from 0.5 to 2, both 1 by default. A change made while reading applies from the next part read. The voice, speed and pitch are remembered on this device.
  • Time — the estimate assumes 183 words a minute, the average rate of reading aloud the Word & Character Counter also uses, divided by the speed. Voices differ, so it is a guide.

Text is read a sentence at a time, so every browser can highlight it. A sentence longer than about ten seconds of speech (150 characters at 1×, up to 300) is read in parts, split at its last comma or space before the limit. Sentences end at full stops, question and exclamation marks, line breaks, and the Devanagari danda (।).

Which voices keep your text private

The browser tells a page whether each voice runs on the device or on a remote service (the Web Speech API's localService), and the voice list shows it. With a voice on your device, the text stays in your browser. With an online voice, the browser sends the text to that voice's speech service:

  • Chrome's voices named “Google …” send it to Google's speech service.
  • Edge's voices with “Online” in their names send it to Microsoft's Azure speech service, which Microsoft says deletes the text and the audio once the speech is made.
  • Any other online voice sends it to its provider, which its name does not state.

Voices on the device come from the operating system, and more can be added:

  • Windows — in Speech settings, choose Add voices under Manage voices and pick a language. Microsoft lists Hemant and Kalpana for Hindi (India), and Ravi and Heera for English (India).
  • Mac — System Settings, Accessibility, Read & Speak: the Info button beside System voice lists the voices to download.
  • Android — Settings, Accessibility, Text-to-speech output: choose the engine and language, and Install voice data for another language.

Sources

Reviewed

Frequently asked questions

Not here. The browser plays its voices straight to the speakers, and the Web Speech API gives the page no recording of them.
The voices are the device's and the browser's own: the operating system's voices, plus online voices some browsers add, such as Chrome's Google voices and Edge's Microsoft voices. The same page lists different voices in another browser or on another device.
Yes, with a Hindi voice. Text in Devanagari picks one automatically when the browser has one; on Windows, add the Hindi voices in Speech settings. One voice reads the whole text, so for Hindi mixed with English, pick the voice for the language most of it is in.
Some voices do not report which word they are on. The sentence being read is highlighted with every voice.

See also