snips-webbrowser-audioserver v1.0.2

snips-webbrowser-audioserver

snips audioserver implementation in a browser with micropohone component for react

About

This package provides a React component that shows a microphone and streams audio over mqtt in a dialog suitable for the Snips Voice Platform making it easy to develop web applications that behave as satellites in the Snips ecosystem.

The package only makes sense when used with a Snips voice server.

Features

• implements audioserver elements of the Snips hermes mqtt protocol supporting streaming audio to and from the device.
• optionally implements the hotword server elements of the Snips hermes mqtt protocol using Porcupine running with WebAssembly in the browser.
• implements the tts server elements of the Snips hermes mqtt protocol using native voices or falling back to speak.js javascript tts generation.
• silence detection by hark so you only transmit while you talk
• long press or right click to show configuration page to select volume, tts voice, hotword, remote control and silence detection.
• logs showing asr transcripts, intent and tts plus audio recordings for each asr transcript.

Screenshots

Install

npm install --save snips-webbrowser-audioserver

!! ensure global script files are included as per example

Usage

import React, { Component } from 'react'

import SnipsMicrophone from 'snips-webbrowser-audioserver'

class Example extends Component {
  render () {
    return (
      <SnipsMicrophone  />
    )
  }
}

Props

• mqttServer
• mqttPort
• hotwordId
• siteId
• clientId
• buttonStyle
• speechBubbleStyle
• configStyle

Config

• inputvolume
• outputvolume
• voicevolume
• ttsvoice
• voicerate
• voicepitch
• remotecontrol
• hotword
• hotwordsensitivity
• silencedetection
• silencesensitivity
• enabletts
• enableaudio
• enablenotifications

License

MIT © syntithenai