Text to speech

What this tool does

Paste or type any text and press Read aloud. The tool uses the speech synthesis built into your browser (the Web Speech API), so the voices are the ones your device and browser provide: the same text can sound quite different on an iPhone, a Windows laptop and an Android phone. You choose the language and the voice, set the speed, pitch and volume, pause and resume, and, where the voice reports its position, see the word being spoken highlighted. Long text is read sentence by sentence so the engine does not stall halfway.

How to use it

1. Paste your text. Up to 50,000 characters at a time; split a whole book into chapters.
2. Pick a language and a voice. The list shows the voices this browser found on your device, your page language first. A voice marked online is synthesised by a remote service, which means the text is sent to that service (Google for Chrome's "Google" voices, Microsoft for Edge's "Online (Natural)" voices). Choose a voice marked on device if the text is private.
3. Set the speed and pitch. 1.0 is the voice's normal speed. Between 1.2 and 1.5 is comfortable for proofreading; slower helps language learners.
4. Read aloud, pause, resume or stop. The status line says which part is being read.

Why the voices differ from device to device

The Web Speech API does not ship voices; it asks the operating system and the browser which ones they have. macOS and iOS offer many voices, some only after you download them in the system settings. Windows has a few built in and more as optional speech packs. Chrome on a computer adds a set of Google voices that work through Google's servers; Edge adds Microsoft's online voices. Android uses whatever text-to-speech engine is installed. If you see no Vietnamese voice, install one in your system's language or speech settings and reload the page. More about this, with what to check on each system, is in our article on how text-to-speech voices differ by device.

Why there is no download button

Browsers play speech synthesis straight to the speakers and do not give the audio to the web page, so no website using this API can save it as an MP3 honestly. Sites that offer a download send your text to their own server and synthesise it there. If you need a file, you can record the computer's sound with a separate app, or record a voice-over yourself with the voice recorder and tidy it up with the audio trimmer.

Tips

• Highlighting depends on the voice: many on-device voices send word positions, some online voices send none. When nothing is highlighted, the tool says so.
• Proofreading by ear catches missing words and repeated words that the eye skips. Paste a draft, set the speed to 1.2 and follow along.
• Numbers and abbreviations are read according to the voice's language: "1.000" is one thousand to a Vietnamese voice and one to an English one. Pick a voice in the text's language.
• Clean up first: the text toolkit can remove line breaks and extra spaces that make a voice pause in odd places.