Speech to text (MERN Stack)

Application that allows users to upload/record speech/conversation of up to 60 minutes, translates the audio and sends a text transcript to the user’s email upon completion.

This project uses the following technologies:

• MongoDB for the database
• Express together with Node for the backend
• React and React Router for frontend
• Google Speech To Text and Google Cloud Storage
• Jest and Enzyme for testing
• Passport and JWTs for authentication
• Socket.io for handling streaming.
• Eslint for clean coding style.

Features

• Register / Login / Logout using JWTs for authentication.
• Allow upload / record audio up to 60 minutes. You can even upload an audio and record your speech at the same time.
• Support multi languages and punctuation.
• Transcribe audio recorded by microphone in real time. For audio file, handle 2 channels audio and support LINEAR16 / FLAC encoding.
• Sending email after transcribing.
• Store user previous transcripts.
• Testing: Snapshot, Unit, Interaction, Mock, Database.

Configuration

Make sure to add your own MONGOURI from your mLab database and your email credentials (to send email using nodemailer) in config/keys.js.

module.exports = {
  mongoURI: "YOUR_MONGO_URI_HERE",
  secretOrKey: "secret",
  username: Your email username,
  password: Your email password,
};

Also, you need your set the environment variable GOOGLE_APPLICATION_CREDENTIALS to the file path of the JSON file that contains your service account key.

export GOOGLE_APPLICATION_CREDENTIALS="./config/credentials.json"

Quick Start

// Install dependencies for server & client
npm install && npm run client-install

// Run client & server with concurrently
npm run dev

// Run test on client & server concurrently
npm run dev-test

// Server runs on http://localhost:5000 and client on http://localhost:3000

Live version can be found here career-foundry.herokuapp.com

You can find sample audio files in dist folder.