Description: Use the chrome.tts API to play synthesized text-to-speech (TTS). See also the related ttsEngine API, which allows an extension to implement a speech engine.
Availability: Since Chrome 36.
Permissions: "tts"
Learn More: Chrome Office Hours: Text to Speech API


Chrome provides native support for speech on Windows (using SAPI 5), Mac OS X, and Chrome OS, using speech synthesis capabilities provided by the operating system. On all platforms, the user can install extensions that register themselves as alternative speech engines.

Generating speech

Call speak() from your extension or Chrome App to speak. For example:

chrome.tts.speak('Hello, world.');

To stop speaking immediately, just call stop():


You can provide options that control various properties of the speech, such as its rate, pitch, and more. For example:

chrome.tts.speak('Hello, world.', {'rate': 2.0});

It's also a good idea to specify the language so that a synthesizer supporting that language (and regional dialect, if applicable) is chosen.

          'Hello, world.', {'lang': 'en-US', 'rate': 2.0});

By default, each call to speak() interrupts any ongoing speech and speaks immediately. To determine if a call would be interrupting anything, you can call isSpeaking(). In addition, you can use the enqueue option to cause this utterance to be added to a queue of utterances that will be spoken when the current utterance has finished.

          'Speak this first.');
          'Speak this next, when the first sentence is done.', {'enqueue': true});

A complete description of all options can be found in the tts.speak below. Not all speech engines will support all options.

To catch errors and make sure you're calling speak() correctly, pass a callback function that takes no arguments. Inside the callback, check runtime.lastError to see if there were any errors.

          function() {
            if (chrome.runtime.lastError) {
              console.log('Error: ' + chrome.runtime.lastError.message);

The callback returns right away, before the engine has started generating speech. The purpose of the callback is to alert you to syntax errors in your use of the TTS API, not to catch all possible errors that might occur in the process of synthesizing and outputting speech. To catch these errors too, you need to use an event listener, described below.

Listening to events

To get more real-time information about the status of synthesized speech, pass an event listener in the options to speak(), like this:

            onEvent: function(event) {
              console.log('Event ' + event.type + ' at position ' + event.charIndex);
              if (event.type == 'error') {
                console.log('Error: ' + event.errorMessage);

Each event includes an event type, the character index of the current speech relative to the utterance, and for error events, an optional error message. The event types are:

  • 'start': The engine has started speaking the utterance.
  • 'word': A word boundary was reached. Use event.charIndex to determine the current speech position.
  • 'sentence': A sentence boundary was reached. Use event.charIndex to determine the current speech position.
  • 'marker': An SSML marker was reached. Use event.charIndex to determine the current speech position.
  • 'end': The engine has finished speaking the utterance.
  • 'interrupted': This utterance was interrupted by another call to speak() or stop() and did not finish.
  • 'cancelled': This utterance was queued, but then cancelled by another call to speak() or stop() and never began to speak at all.
  • 'error': An engine-specific error occurred and this utterance cannot be spoken. Check event.errorMessage for details.

Four of the event types—'end', 'interrupted', 'cancelled', and 'error'—are final. After one of those events is received, this utterance will no longer speak and no new events from this utterance will be received.

Some voices may not support all event types, and some voices may not send any events at all. If you do not want to use a voice unless it sends certain events, pass the events you require in the requiredEventTypes member of the options object, or use getVoices() to choose a voice that meets your requirements. Both are documented below.

SSML markup

Utterances used in this API may include markup using the Speech Synthesis Markup Language (SSML). If you use SSML, the first argument to speak() should be a complete SSML document with an XML header and a top-level <speak> tag, not a document fragment.

For example:

          '<?xml version="1.0"?>' +
          '<speak>' +
          '  The <emphasis>second</emphasis> ' +
          '  word of this sentence was emphasized.' +

Not all speech engines will support all SSML tags, and some may not support SSML at all, but all engines are required to ignore any SSML they don't support and to still speak the underlying text.

Choosing a voice

By default, Chrome chooses the most appropriate voice for each utterance you want to speak, based on the language. On most Windows, Mac OS X, and Chrome OS systems, speech synthesis provided by the operating system should be able to speak any text in at least one language. Some users may have a variety of voices available, though, from their operating system and from speech engines implemented by other Chrome extensions. In those cases, you can implement custom code to choose the appropriate voice, or to present the user with a list of choices.

To get a list of all voices, call getVoices() and pass it a function that receives an array of TtsVoice objects as its argument:

          function(voices) {
            for (var i = 0; i < voices.length; i++) {
              console.log('Voice ' + i + ':');
              console.log('  name: ' + voices[i].voiceName);
              console.log('  lang: ' + voices[i].lang);
              console.log('  extension id: ' + voices[i].extensionId);
              console.log('  event types: ' + voices[i].eventTypes);


speak chrome.tts.speak(string utterance, TtsOptions options, function callback)
stop chrome.tts.stop()
pause chrome.tts.pause()
resume chrome.tts.resume()
isSpeaking chrome.tts.isSpeaking(function callback)
getVoices chrome.tts.getVoices(function callback)



"start", "end", "word", "sentence", "marker", "interrupted", "cancelled", "error", "pause", or "resume"


"male", or "female"


Since Chrome 77.

The speech options for the TTS engine.
boolean (optional) enqueue

If true, enqueues this utterance if TTS is already in progress. If false (the default), interrupts any current speech and flushes the speech queue before speaking this new utterance.

string (optional) voiceName

The name of the voice to use for synthesis. If empty, uses any available voice.

string (optional) extensionId

The extension ID of the speech engine to use, if known.

string (optional) lang

The language to be used for synthesis, in the form language-region. Examples: 'en', 'en-US', 'en-GB', 'zh-CN'.

VoiceGender (optional) gender

Deprecated since Chrome 77. Gender is deprecated and will be ignored.

Gender of voice for synthesized speech.

double (optional) rate

Speaking rate relative to the default rate for this voice. 1.0 is the default rate, normally around 180 to 220 words per minute. 2.0 is twice as fast, and 0.5 is half as fast. Values below 0.1 or above 10.0 are strictly disallowed, but many voices will constrain the minimum and maximum rates further—for example a particular voice may not actually speak faster than 3 times normal even if you specify a value larger than 3.0.

double (optional) pitch

Speaking pitch between 0 and 2 inclusive, with 0 being lowest and 2 being highest. 1.0 corresponds to a voice's default pitch.

double (optional) volume

Speaking volume between 0 and 1 inclusive, with 0 being lowest and 1 being highest, with a default of 1.0.

array of string (optional) requiredEventTypes

The TTS event types the voice must support.

array of string (optional) desiredEventTypes

The TTS event types that you are interested in listening to. If missing, all event types may be sent.

function (optional) onEvent

This function is called with events that occur in the process of speaking the utterance.

TtsEvent event

The update event from the text-to-speech engine indicating the status of this utterance.


An event from the TTS engine to communicate the status of an utterance.
EventType type

The type can be start as soon as speech has started, word when a word boundary is reached, sentence when a sentence boundary is reached, marker when an SSML mark element is reached, end when the end of the utterance is reached, interrupted when the utterance is stopped or interrupted before reaching the end, cancelled when it's removed from the queue before ever being synthesized, or error when any other error occurs. When pausing speech, a pause event is fired if a particular utterance is paused in the middle, and resume if an utterance resumes speech. Note that pause and resume events may not fire if speech is paused in-between utterances.

integer (optional) charIndex

The index of the current character in the utterance. For word events, the event fires at the end of one word and before the beginning of the next. The charIndex represents a point in the text at the beginning of the next word to be spoken.

string (optional) errorMessage

The error description, if the event type is error.

integer (optional) length

Since Chrome 74.

The length of the next part of the utterance. For example, in a word event, this is the length of the word which will be spoken next. It will be set to -1 if not set by the speech engine.


A description of a voice available for speech synthesis.
string (optional) voiceName

The name of the voice.

string (optional) lang

The language that this voice supports, in the form language-region. Examples: 'en', 'en-US', 'en-GB', 'zh-CN'.

VoiceGender (optional) gender

Deprecated since Chrome 70. Gender is deprecated and will be ignored.

This voice's gender.

boolean (optional) remote

If true, the synthesis engine is a remote network resource. It may be higher latency and may incur bandwidth costs.

string (optional) extensionId

The ID of the extension providing this voice.

array of EventType (optional) eventTypes

All of the callback event types that this voice is capable of sending.



chrome.tts.speak(string utterance, TtsOptions options, function callback)

Speaks text using a text-to-speech engine.

string utterance

The text to speak, either plain text or a complete, well-formed SSML document. Speech engines that do not support SSML will strip away the tags and speak the text. The maximum length of the text is 32,768 characters.

TtsOptions (optional) options

The speech options.

function (optional) callback

Called right away, before speech finishes. Check runtime.lastError to make sure there were no errors. Use options.onEvent to get more detailed feedback.

If you specify the callback parameter, it should be a function that looks like this:

function() {...};



Stops any current speech and flushes the queue of any pending utterances. In addition, if speech was paused, it will now be un-paused for the next call to speak.



Pauses speech synthesis, potentially in the middle of an utterance. A call to resume or stop will un-pause speech.



If speech was paused, resumes speaking where it left off.


chrome.tts.isSpeaking(function callback)

Checks whether the engine is currently speaking. On Mac OS X, the result is true whenever the system speech engine is speaking, even if the speech wasn't initiated by Chrome.

function (optional) callback

If you specify the callback parameter, it should be a function that looks like this:

function(boolean speaking) {...};
boolean speaking

True if speaking, false otherwise.


chrome.tts.getVoices(function callback)

Gets an array of all available voices.

function (optional) callback

If you specify the callback parameter, it should be a function that looks like this:

function(array of TtsVoice voices) {...};
array of TtsVoice voices

Array of tts.TtsVoice objects representing the available voices for speech synthesis.