11Labs Clone - AI Voice Generation Platform

This project is a comprehensive AI voice generation platform that combines multiple state-of-the-art text-to-speech and voice cloning technologies. The platform includes a modern Next.js frontend and multiple AI model backends for different voice generation capabilities.

Features

• Text-to-Speech with StyleTTS2
• Voice Cloning with SEED-VC
• Audio Generation with Make-An-Audio
• Modern Next.js Frontend with TypeScript
• Docker-based deployment
• GPU-accelerated AI models
• User authentication and management
• File storage with AWS S3

Prerequisites

• Node.js 18+ and pnpm 9.12.1
• Docker and Docker Compose
• NVIDIA GPU with CUDA support
• AWS Account (for S3 storage)
• PostgreSQL database

Installation

1. Clone the repository:

git clone https://github.com/khaireddine-arbouch/11labs-clone.git
cd 11labs-clone

2. Install frontend dependencies:

cd 11labs-clone-frontend
pnpm install

3. Set up environment variables: Create a .env file in the frontend directory with:

DATABASE_URL="postgresql://user:password@localhost:5432/dbname"
NEXTAUTH_SECRET="your-secret"
NEXTAUTH_URL="http://localhost:3000"
AWS_ACCESS_KEY_ID="your-aws-key"
AWS_SECRET_ACCESS_KEY="your-aws-secret"
AWS_REGION="your-aws-region"
AWS_BUCKET_NAME="your-bucket-name"

4. Initialize the database:

pnpm db:generate
pnpm db:push

5. Start the development server:

pnpm dev

Docker Deployment

The project uses Docker Compose to manage multiple services:

1. StyleTTS2 API (Port 8000)
2. SEED-VC API (Port 8001)
3. Make-An-Audio API (Port 8002)

To start all services:

docker-compose up -d

Project Structure

.
├── 11labs-clone-frontend/     # Next.js frontend application
│   ├── src/                  # Source code
│   ├── prisma/              # Database schema and migrations
│   └── public/              # Static assets
├── StyleTTS2/               # StyleTTS2 voice generation service
├── seed-vc/                # SEED-VC voice cloning service
├── Make-An-Audio/          # Audio generation service
└── docker-compose.yml      # Docker services configuration

Technology Stack

Frontend

• Next.js 15.2.3
• React 19.0.0
• TypeScript
• TailwindCSS 4.0.15
• tRPC
• Prisma
• NextAuth.js
• Zustand for state management

Backend Services

• StyleTTS2 for text-to-speech
• SEED-VC for voice cloning
• Make-An-Audio for audio generation
• All services are containerized with Docker

API Documentation

StyleTTS2 API

• Endpoint: http://localhost:8000
• GPU-accelerated text-to-speech generation

SEED-VC API

• Endpoint: http://localhost:8001
• Voice cloning capabilities

Make-An-Audio API

• Endpoint: http://localhost:8002
• Audio generation and manipulation

Security

• All API endpoints are protected with authentication
• AWS S3 is used for secure file storage
• Environment variables for sensitive data
• HTTPS enforced in production

Deployment

1. Build the frontend:

cd 11labs-clone-frontend
pnpm build

2. Start the Docker services:

docker-compose up -d

3. The application will be available at:

• Frontend: http://localhost:3000
• StyleTTS2 API: http://localhost:8000
• SEED-VC API: http://localhost:8001
• Make-An-Audio API: http://localhost:8002

Acknowledgments

• StyleTTS2: https://github.com/keithito/tacotron
• SEED-VC: https://github.com/keithito/tacotron
• Make-An-Audio: https://github.com/keithito/tacotron