LexiTalk

Overview

LexiTalk is a language tutor web application built using Vite + React for the frontend and Express.js for the backend. The app utilizes LiveKit Cloud for real-time audio streaming and ElevenLabs for text-to-speech conversion. An LLM API is integrated to generate language learning responses.

Features

• Real-time audio streaming using LiveKit
• Text-to-speech conversion via ElevenLabs
• Chat interaction powered by an LLM
• Audio recording and playback
• User authentication and session management
• Secure API with rate limiting and error handling

---

Tech Stack

Frontend:

• React (Vite)
• React Router
• Context API / Hooks
• LiveKit SDK
• Tailwind CSS (for styling)

Backend:

• Node.js + Express
• LiveKit Server SDK
• ElevenLabs API
• LLM Service (OpenAI / Anthropic / etc.)
• MongoDB / PostgreSQL (for user data)
• JWT Authentication

---

Project Structure

Frontend

src/
├── components/
│   ├── Chat/
│   │   ├── Chat.jsx
│   │   ├── ChatMessage.jsx
│   │   └── ChatInput.jsx
│   ├── Audio/
│   │   ├── AudioRecorder.jsx
│   │   └── AudioPlayer.jsx
│   └── LiveKit/
│       └── LiveKitProvider.jsx
├── services/
│   ├── api.js         # Backend API calls
│   ├── livekit.js     # LiveKit service functions
│   └── elevenlabs.js  # ElevenLabs integration
├── hooks/
│   ├── useChat.js
│   └── useAudio.js
├── context/
│   ├── ChatContext.jsx
│   └── AudioContext.jsx
├── App.jsx
└── main.jsx

Backend

backend/
├── config/
│   ├── default.js     # Configuration variables
│   └── livekit.js     # LiveKit specific config
├── controllers/
│   ├── chatController.js
│   ├── audioController.js
│   └── userController.js
├── middleware/
│   ├── auth.js
│   ├── errorHandler.js
│   └── rateLimiter.js
├── routes/
│   ├── chatRoutes.js
│   ├── audioRoutes.js
│   └── userRoutes.js
├── services/
│   ├── llmService.js  # Integration with your chosen LLM
│   ├── elevenlabsService.js
│   └── livekitService.js
├── utils/
│   ├── logger.js
│   └── responseFormatter.js
├── app.js            # Express application setup
└── server.js         # Server entry point

---

Installation & Setup

Prerequisites

• Node.js & npm
• MongoDB / PostgreSQL (if using a database)
• LiveKit Cloud Account
• ElevenLabs API Key
• LLM API Key (OpenAI / Anthropic / etc.)

Backend Setup

cd backend
npm install
cp .env.example .env # Add required API keys in .env
node server.js

Frontend Setup

cd frontend
npm install
npm run dev

---

API Endpoints

Auth Routes

Method	Endpoint	Description
POST	/api/auth/register	Register a new user
POST	/api/auth/login	User login

Chat Routes

Method	Endpoint	Description
POST	/api/chat	Send message to LLM

Audio Routes

Method	Endpoint	Description
POST	/api/audio/speak	Convert text to speech (ElevenLabs)
POST	/api/audio/stream	Handle LiveKit audio streaming

---

Contributing

Feel free to fork this repository and submit a pull request with improvements or bug fixes.

---

License

This project is licensed under the MIT License.