Web Speech Recognizer

A convenient client for Web Speech Recognition API.

In addition to handle any interaction with the Web Speech Recognition API supported by HTML 5, it notifies when the user speaks and what's the volume of the sound.

Requirements

Web Speech Recognition API is not supported by all the browsers. Here you can check the browser support https://caniuse.com/#feat=speech-recognition

Install

npm install web-speech-recognizer

Usage

import WebSpeechRecognizer from 'web-speech-recognizer';

const speechRecognizer = WebSpeechRecognizer.init({
  lang: 'en-US',
  continuesRecognition: true,
  onUserSpeech: (recognition) => {
    console.log('Speech recognition', recognition)
  },
  onUserSpeak: (volume) => {
    console.log('Speak volume', volume)
  }
})

API

Requiring the module gives you an object that defines one primary function.

init(settings)

Useful to initialize the speech recognition service.

Arguments

(Object): An object with the following properties:

• lang (String): Language code (ISO 639-1). Default to browser language. i.e. "en-US"
• continuesRecognition (Bool): When continues recognition mode is turned on, the service will be continuously listening and returning transcriptions for what was received. If the continues recognition mode is turned off, the service will start listening after calling the start method, and it will stop listening as soon as the first transcription result has been reported or the stop method is called. Default to true.
• onUserSpeech (function): A callback function useful to handle the status of the speech recognition service and the speech transcriptions.
• onUserSpeak (function): A callback function to handle the user speak volume (0 - 100).

Returns

(Object): Speech recognizer client that defines two methods start and stop.

Examples

Continues recognition

import WebSpeechRecognizer from 'web-speech-recognizer';

const speechRecognizer = WebSpeechRecognizer.init({
  lang: 'it-IT',
  continuesRecognition: true,
  onUserSpeech: (recognition) => {
    console.log('Speech recognition', recognition)
  },
  onUserSpeak: (volume) => {
    console.log('Speak volume', volume)
  }
})

speechRecognizer.start();
// Now it is listening
// You can stop the service executing `speechRecognizer.stop()` and resume it executing `speechRecognizer.start()`

On demand recognition

import WebSpeechRecognizer from 'web-speech-recognizer';

const speechRecognizer = WebSpeechRecognizer.init({
  lang: 'it-IT',
  continuesRecognition: false,
  onUserSpeech: (recognition) => {
    console.log('Speech recognition', recognition)
  },
  onUserSpeak: (volume) => {
    console.log('Speak volume', volume)
  }
})

speechRecognizer.start();
// Now it is listening
// It will stop listening as soon as the first final results has been reported to the callback defined in `onUserSpeech`.
// As second option, you can stop the service executing `speechRecognizer.stop()`
// If you want to resume the speech recognition you have to re-execute `speechRecognizer.start()`