Electron-vosk-speech

Lightweight Speech Recognition Library for Electron. Based on nodejs-speech-kiosk-usercase and vosk-api.

1. Install:

yarn add electron-vosk-speech

Then you need to install docker.

Run vosk-api local server:
sudo docker run -d -p 2700:2700 alphacep/kaldi-ru:latest
docker run -d -p 2700:2700 alphacep/kaldi-ru:latest

To restart:
docker restart alphacep/kaldi-ru:latest
To stop and clear vosk container:
docker stop vosk

To clear all docker containers:
``
docker stop $(docker ps -aq)
docker rm $(docker ps -aq)
`








Example usage:


Add this to Electron's render process:

`
const {Recognizer} = require('electron-vosk-speech')
const {ipcRenderer} = require('electron')

const Rec = new Recognizer({
ipcRenderer,
onSpeechRecognized: res => console.log('РЕЗУЛЬТАТ! ' + JSON.stringify(res)),
onSpeechStart: () => console.log('ГОВОРИТ!'), // fires on speech started
onSpeechEnd: () => console.log('ЗАМОЛЧАЛ!'), // fires on speech ended
options: {
isSpeech2Text: true,
autoInit: true,
forced: true, // forced start recording
idleDelay: 5000,
languageCode: 'ru',
harkOptions: {},
gsFormat: false // if true, returns result in GoogleSpeech format
// for backward compatibility with solutions based on GoogleSpeech
}
})

// Rec.startAll() - start listening, recording and recognize
// Rec.stopAll() - stop listening, recording and recognize

// Rec.startListening() - start listening
// Rec.stopListening() - stop listening

// Rec.stopRecognize() - stop recording and recognize
// Rec.startRecognize() - start recording and recognize
`

Add this to Electron's main process:
`
const { app, BrowserWindow } = require('electron')
const {speechSaverHandler, connect2Vosk } = require('electron-vosk-speech/src/utils')

const win = new BrowserWindow({
width: 800,
height: 600,
webPreferences: {
nodeIntegration: true
}
})

// default values


const {VoskConnector} = require('electron-vosk-speech/src/utils')
const vosk = new VoskConnector({
// sudo: true,
autostart: 1,
sudo: 0,
docker: {
image: 'neurocity/vosk-small-ru', //todo добавить ошибку, или убивание процесса, если данные имена не совпадают
version: '0.0.2'
},
config: {
word_list: words
}
})
vosk.init(mainWindow.webContents, (ws) => {
mainWindow.webContents.session.on('will-download', function voskSpeechSaver(...rest){
vosk.speechSaverHandler(app.getAppPath(), ws, ...rest) // for new version
})
})
``
Enjoy!