speech

Speech APIs for real time voice analysis.

This is an experiment in combining, speech to text with tex to speech and language generation.

You are super welcome to help out. <3

How to run

After cloning the repo you need the react server, python server and text-to-speech docker server to all be running. Follow the steps for each to get it working on your machine.

React web app

npm install
npm start

React web app fix

export NODE_OPTIONS=--openssl-legacy-provider

Python GPT2 server

cd python
bash run_local.sh

This spins up a local server running GPT2 hosted on FastApi.

Text to speech docker

https://github.com/synesthesiam/docker-mozillatts

docker run -it -p 5002:5002 synesthesiam/mozillatts

Speech to text web

https://www.npmjs.com/package/react-speech-recognition

import React from 'react'
import SpeechRecognition, { useSpeechRecognition } from 'react-speech-recognition'
import { playAudio } from './App.js'

const Dictaphone = () => {
  const { transcript, resetTranscript } = useSpeechRecognition()

  if (!SpeechRecognition.browserSupportsSpeechRecognition()) {
    return null
  }

  return (
    <div>
      <button onClick={SpeechRecognition.startListening}>Start</button>
      <button onClick={SpeechRecognition.stopListening}>Stop</button>
      <button onClick={resetTranscript}>Reset</button>
      <button onClick={() => playAudio(transcript)}>Transcribe</button>
      <p>{transcript}</p>
    </div>
  )
}
export default Dictaphone

Running inside react

export async function playAudio(text) {
  var audio = new Audio(`http://localhost:5002/api/tts?text=${encodeURIComponent(text)}`);  
  audio.type = 'audio/wav';

  try {
    await audio.play();
    console.log('Playing...');
  } catch (err) {
    console.log('Failed to play...' + err);
  }
}