Acente

Introduction

This project is apart of the Fall 2021 GDSC Community Projects.

Acente, a web application designed to be a personalized accent coach for users. It will use Speech Recognition to provide actionable feedback to improve a user’s accent. The web app will also store and manage a user’s practice sentences, in order to create a curated accent profile that the user can use to work on their weak spots. The user will also have access to a sandbox, to try out custom sentences that they want to improve on. The website will provide actionable feedback through the use of IPA (international Phonetic Alphabet), by breaking down the sentence.

Tech Stack

For the Backend:

• Google Speech to Text
• Flask
• Python

For the Frontend:

• Javascript
• React
• Firebase Auth & Hosting

Directory Structure

The backend code can be found in backend/ and the frontend code can be found in frontend/

Frontend Files

• Dashboard.js: Displays the user's accent profile. User can view their progress, strengths, weaknesses, and custom sentences they have created.
• Footer.js: Displays the Acente Footer.
• NavBar.js: Adaptive Navigation bar displays login and sign-up buttons if user has not logged in. After authentication, it displays the burger menu.
• Sandbox.js: User can create and practice on custom sentences.
• Test.js: User can practice on predefined sentences.

Secrets

Backend

The backend requires 3 keys:

• fireStoreKey.json, a GCP key to the FireStore database
• service-key-account.json, a GCP service account key
• secrets.json, a GCP key to the real time Firebase database

Frontend

The frontend has an .env file with:

• REACT_APP_BACKEND_URL, the url to the backend