Text to Speech
Text to speech reads your text aloud using the Web Speech API and whatever voices are installed on your device. Nothing is sent to a server, which also means the available voices and their quality depend entirely on your operating system rather than on this page.
How it works
The browser exposes the speech synthesis engine already present in the operating system. macOS and iOS ship high-quality voices, Windows includes several, and Android varies by manufacturer. Some Chrome voices are cloud-backed and need a connection.
- Rate runs from 0.5x to 2x. Values above about 1.6 become hard to follow on most voices.
- Pitch runs from 0 to 2, where 1 is the voice’s natural pitch.
- The voice list loads asynchronously in Chrome, so it can be empty for a moment after the page opens.
- Voices are labelled with their language tag, which is what determines pronunciation rules.
Choosing a voice whose language does not match the text produces the wrong pronunciation rules. English text read by a German voice is a reliable way to hear this.
Examples
Reading at a natural pace
Rate
1.0x
Pitch
1.0
Result
The text is spoken in the selected voice at its natural speed
A rate of 1 is the voice’s own default, which varies between voices. Some are naturally faster than others.
Skim-reading speed
Rate
1.6x
Result
Noticeably faster, still intelligible on most voices
Beyond about 1.6 the synthesis quality degrades on lower-quality voices, and consonants start to run together.
Frequently asked questions
Why do I have different voices from someone else?
Because the voices come from your operating system, not from this page. macOS and iOS ship a large set including several high-quality ones; Windows includes fewer; Android depends on the manufacturer and the installed speech engine. The tool lists whatever your device provides.
Can I download the audio as a file?
No. The Web Speech API plays audio through the system and provides no way to capture it as a file. The specification simply does not expose the audio stream. Recording it requires screen or system audio capture outside the browser.
Why is the voice list empty when the page first loads?
Chrome loads voices asynchronously, so the first query often returns nothing. The tool listens for the voiceschanged event and refreshes the list when they arrive, which usually takes under a second.
Does the text get sent anywhere?
Not for most voices, which run locally on your device. Some Chrome voices are cloud-backed and do send text to Google for synthesis. Those are marked in the voice list by their naming, and choosing a local voice avoids it entirely.