🎵 VerseForge - AI Music Generation Platform

The world's most advanced AI music creation platform. Create professional songs, beats, and compositions with cutting-edge artificial intelligence.

🌟 Overview

VerseForge is a cutting-edge AI-powered music generation platform that transforms your creative ideas into professional-quality songs. Using advanced artificial intelligence, it generates original lyrics and converts them into high-quality audio tracks with customizable parameters.

✨ Key Features

• 🎤 AI Lyrics Generation - Create original, artist-style lyrics using advanced language models
• 🎵 Text-to-Speech Audio - Convert lyrics into professional-quality audio tracks
• 🎨 Custom Album Covers - Upload and customize album artwork for each song variant
• 🎛️ Advanced Controls - Fine-tune creativity, vocal balance, and tempo
• 👤 User Authentication - Secure user management with Clerk
• 📱 Responsive Design - Beautiful, modern UI that works on all devices
• 🎧 Audio Player - Full-featured player with lyrics display and waveform visualization
• 🔄 Multiple Variants - Generate 2 different versions of each song
• 💾 Session Persistence - Your creations persist across page refreshes

🚀 Quick Start

Prerequisites

• Node.js (v18.0.0 or higher)
• npm or yarn
• API Keys (see Environment Variables section)

Installation & Setup

Step 1: Clone the Repository

# Clone the repository
git clone https://github.com/Nixxx19/verseforge.git
cd verseforge

Step 2: Install Dependencies

# Install frontend dependencies
npm install

# Install backend dependencies
cd backend
npm install

Step 3: Start the Application

# Start backend server
cd backend
npm start

In a new terminal window:

# Start frontend server
npm run dev

🔬 How It Works

VerseForge uses a sophisticated multi-stage AI pipeline to transform your creative ideas into professional-quality music. Here's the technical breakdown of our advanced system:

Stage 1: Custom LSTM Preprocessing

Before generating lyrics, we run your input through our custom LSTM (Long Short-Term Memory) neural network that has been specifically trained on music data:

• Purpose: Preprocesses and enhances your input to create "cloud words" and cleaned lyrics
• Training Data: Trained on extensive music datasets to understand lyrical patterns and structures
• Output: Refined, music-optimized input that's ready for the next stage

Stage 2: Fine-Tuned Llama 3.3 Model

Our core lyrics generation is powered by Meta's open-source Llama 3.3 model, which we've fine-tuned specifically for music:

• Base Model: meta-llama/llama-3.3-8b-instruct:free
• Fine-Tuning: Custom fine-tuning on music-specific datasets
• Specialization: Trained to understand artist styles, lyrical structures, and musical themes
• Output: High-quality, artist-style lyrics that match your creative vision

Stage 3: Advanced TTS with Custom Tags

The generated lyrics are then processed by Sunauto TTS, which we've enhanced with custom fine-tuning:

• Custom Tags: Created 1000+ specialized tags for better audio output
• Fine-Tuning: Custom training to optimize for music generation
• Quality Enhancement: Improved vocal synthesis and musical expression
• Output: Professional-quality audio tracks with natural-sounding vocals

Technical Pipeline Flow

User Input → LSTM Preprocessing → Llama 3.3 Generation → Sunauto TTS → Final Audio
      ↓              ↓                    ↓                ↓
Eg: "Drake song"  → Cloud Words → Artist-style Lyrics → Professional Audio

Why This Approach Works

1. LSTM Preprocessing: Ensures input is optimized for music generation
2. Fine-Tuned Llama: Leverages open-source power with music specialization
3. Custom TTS Tags: 1000+ tags provide precise control over audio output
4. End-to-End Pipeline: Each stage is optimized for the next, ensuring quality

This multi-stage approach allows us to generate music that's both creative and technically excellent, combining the power of open-source models with our specialized fine-tuning.

🎵 How to Use

1. Generate Your First Song

1. Navigate to the homepage - You'll see the main interface with a prompt input
2. Enter your musical vision - Describe the song you want to create:
   • "Drake song about late night vibes"
   • "Taylor Swift ballad about heartbreak"
   • "Travis Scott trap song about success"
3. Adjust settings (optional):
   • Temperature (1.5-2.0): Controls lyrical creativity and originality
   • Balance (0-2.0): Controls vocal balance and instrumental focus
   • BPM (120-300): Sets the tempo and energy of the track
4. Click "Generate" - The system will create your song

2. Customize Your Song

• Edit Song Name: Click on the song title to rename it
• Upload Custom Cover: Hover over album art and click "Upload Cover"
• Switch Variants: Use arrow buttons to explore different versions
• Play Controls: Use the full-featured audio player with lyrics display

3. Explore Features

• Lyrics Display: Click the lyrics button to see synchronized lyrics
• Waveform Visualization: Watch the animated waveform during playback
• Volume Control: Adjust volume or mute the audio
• Progress Bar: Click to seek to any point in the song

🏗️ Project Architecture

Frontend (React + TypeScript)

src/
├── components/           # React components
│   ├── ui/              # Reusable UI components (shadcn/ui)
│   ├── Header.tsx       # Navigation header
│   ├── HeroSection.tsx  # Main landing section
│   ├── GeneratePage.tsx # Song generation page
│   ├── AudioPlayerSection.tsx # Audio showcase
│   ├── FeaturesSection.tsx    # Features showcase
│   └── ...
├── pages/               # Page components
│   ├── Index.tsx        # Homepage
│   └── NotFound.tsx     # 404 page
├── contexts/            # React contexts
│   └── AuthContext.tsx  # Authentication context
├── hooks/               # Custom React hooks
├── lib/                 # Utility functions
└── main.tsx            # Application entry point

Backend (Express.js + Node.js)

backend/
├── server.js           # Main server file
├── package.json        # Backend dependencies
└── .env               # Environment variables

Key Technologies

• Frontend: React 18, TypeScript, Vite, Tailwind CSS
• Backend: Express.js, Node.js
• UI Components: Radix UI, shadcn/ui
• Authentication: Clerk
• Styling: Tailwind CSS with custom design system
• Icons: Lucide React
• State Management: React Hooks, Context API

🎨 Design System

VerseForge features a sophisticated design system with:

Color Palette

• Primary: Gradient blues and purples
• Accent: Warm oranges and reds
• Background: Dark theme with glass morphism
• Text: High contrast white and muted tones

Typography

• Display: Clash Display (headings)
• Body: Inter (main text)
• Monospace: For code and lyrics

Components

• Glass Cards: Frosted glass effect with backdrop blur
• Gradient Buttons: Animated gradient backgrounds
• Custom Shadows: Multi-layered shadow system
• Responsive Grid: Mobile-first responsive desig

Code Structure

• Components: Modular, reusable React components
• Hooks: Custom hooks for state management
• Contexts: Global state management
• Utils: Helper functions and utilities
• Types: TypeScript type definitions

Security & Authentication

• User Authentication: Powered by Clerk
• API Security: CORS enabled for development
• Input Validation: Server-side validation for all inputs

Browser Support

• Chrome: 90+
• Firefox: 88+
• Safari: 14+
• Edge: 90+