ResponsiveVoice Text To Speech API

Include the JS file in your page

speak(string text, [string voice], [object parameters])

Starts speaking the text in a given voice.
text: String
The text to be spoken.
voice: String
Defaults to “UK English Female”. Choose from the available ResponsiveVoices.
parameters: Object
Used to add optional pitch (range 0 to 2), rate (range 0 to 1.5), volume (range 0 to 1) and callbacks.
Pitch, rate and volume may not affect audio on some browser combinations, older versions of Chrome on Windows for example.

Speak a specified element on the page:


Stops playing the speech.


Checks if browser supports native TTS

Returns: true/false


Returns: a list of available voices

setDefaultVoice(string voice)

Allows setting a default voice, which will be used by responsiveVoice.speak whenever a voice is not specified as a parameter.

setDefaultRate(number rate)

Allows setting a default rate of speech, which will be used by responsiveVoice.speak whenever rate is not included in the parameters. The rate parameter must be > 0 and <= 1.5
Please note this might not work with all voices, as certain voices don’t support variable rate of speech.


Detects if native TTS or TTS audio element is producing output.

Returns: true/false

pause() and resume()

Pauses/Resumes speech

setTextReplacements(Array replacements)

Replaces selected words or expressions in the text. Useful for specifying pronunciation variations for different voices.

responsiveVoice.setTextReplacements expects an array of objects where each object is a replacement. The replacement object has the following parameters:

  • searchvalue: (required) text to be replaced. Regular expressions are supported.
  • newvalue: (required) replacement text.
  • collectionvoices: (optional) Voice name (from ResponsiveVoice collection) for which the replacement will be applied. Can be a unique name or an array of names.
  • systemvoices: (optional) Voice name (from System voices collection) for which the replacement will be applied. Can be a unique name or an array of names.


Replace “human” with “robot”. It will be replaced for any voice selected.

Replace “human” with “robot” and “dog” with “cat”.

Replace “human” with “robot” only for “Google UK English Female” voice in responsiveVoice.voicecollection.

Replace “human” with “robot” for “de-DE” and “fr-FR” system voices. These correspond to german and french on iOS devices.

Replace any combination of numbers with the word “numbers”.


On some devices, such as mobile, browsers prevent audio from being played without a user gesture. ResponsiveVoice can listen for a click on the window and take it as the user gesture required by the browser. This will grant ResponsiveVoice permission to play any audio from that moment on.

Note: This click hook is automatically enabled on mobile devices.

Alternatively, in the case that listening for a click on the window is not possible, responsiveVoice.clickEvent() can be called directly from any user gesture and it will grant ResponsiveVoice the required permission.


Sometimes performance can have an impact on the TTS engine, causing unexpected behaviors when the system load is high. ResponsiveVoice uses a timeout as a fallback mechanism so the onstart and onend events will be triggered normally even if the TTS engine didn’t respond to the speak request. This is specially useful for a multiple phrase setup such as a dialog where the next phrase needs to be played after the current one.

Note: enableEstimationTimeout is enabled by default. 

We recommend to disable this feature for non-latin character languages. It can be disabled by setting the parameter as false:

Are You Interested In