🎵 H## Features

• 🎼 Real-time Music Generation: Uses Google's Lyria RealTime model for high-quality instrumental music
• ⏱️ 1-Minute Segments: Perfect length for immersive coding sessions
• 🔄 Parallel Processing: Generates next segment while current segment plays for seamless experience
• 🎵 Customizable Parameters: Set BPM, mood, and music themes
• 🎯 Coding-Focused: Designed specifically for background music while coding
• 🌈 Beautiful CLI Interface: Colorful and interactive command-line experience
• ⚡ No-Gap Playback: Seamless transitions between segments
• 🧹 Smart Memory Management: Automatic cleanup of old segments

An AI-powered music generation CLI tool that creates personalized instrumental music for coding sessions using Google's Gemini Lyria RealTime model.

Features

• 🎼 Real-time Music Generation: Uses Google's Lyria RealTime model for high-quality instrumental music
• 🎛️ Two Generation Modes:
  • Interactive Mode: Generate 15-second segments on demand, perfect for focused work
  • Continuous Mode: Auto-generate 30-second segments for uninterrupted listening
• 🔄 Parallel Processing: Generates next segment while current segment plays for seamless experience
• � Customizable Parameters: Set BPM, mood, and music themes
• � Coding-Focused: Designed specifically for background music while coding
• 🌈 Beautiful CLI Interface: Colorful and interactive command-line experience
• ⚡ No-Gap Playback: Seamless transitions between segments
• 🧹 Smart Memory Management: Automatic cleanup of old segments

Installation

1. Clone or download this project
2. Install dependencies:

   npm install

3. Get your Gemini API key:
   • Visit Google AI Studio
   • Create or use an existing API key
   • Make sure you have access to the Lyria RealTime model

Usage

Run the CLI tool:

npx harmony-cli

Or if installed globally:

npm install -g .
harmony-cli

Interactive Setup

The CLI will guide you through:

1. API Key Input: Enter your Gemini API key
2. Music Theme: Describe the type of music you want (e.g., "chill lo-fi hip hop", "ambient electronic", "peaceful piano")
3. Tempo Selection: Choose BPM from predefined options (70-160 BPM)
4. Mood Selection: Pick from various moods like Chill, Upbeat, Ambient, etc.

How It Works

🔄 Continuous Generation Workflow:

1. Generate first 1-minute segment
2. Play current segment + Generate next segment (in parallel)
3. When both complete, seamlessly move to the new segment
4. Repeat the cycle automatically

Perfect for coding because:

• ✅ 1-minute segments: Ideal length for immersive musical experience
• ✅ No interruptions: Generation happens while music plays
• ✅ No gaps: Seamless transitions between segments
• ✅ Continuous variety: New AI-generated music every minute

Example Prompts

Music Themes:

• "minimal techno for deep focus"
• "jazz fusion with smooth bass"
• "ambient electronic soundscape"
• "lo-fi hip hop with vinyl crackle"
• "classical piano with gentle strings"
• "synthwave retro vibes"

Available Moods:

• Chill
• Upbeat
• Ambient
• Energetic
• Peaceful
• Dreamy
• Focused

How It Works

1. Connection: Establishes connection to Gemini Lyria RealTime via Python bridge
2. Initial Generation: Creates first 30-second music segment
3. Parallel Processing: While playing current segment, generates next segment in background
4. Seamless Playback: Transitions between segments with no gaps
5. Continuous Loop: Keeps generating and playing until user stops with Ctrl+C
6. Smart Cleanup: Automatically removes old segments to save disk space

Parallel Generation Flow

🎼 Generate Segment 1 → 🎵 Play Segment 1
                        ↓ (while playing)
                        🎼 Generate Segment 2 → 🎵 Play Segment 2  
                                               ↓ (while playing)
                                               🎼 Generate Segment 3...

Technical Details

• Audio Format: 16-bit PCM, 48kHz, Stereo
• API: Google Gemini Lyria RealTime via WebSocket
• Latency: Low-latency real-time streaming
• Quality: High-fidelity instrumental music
• Watermarking: All generated audio includes identification watermarks

Configuration Options

The CLI automatically configures:

• BPM: 60-200 (user selectable)
• Temperature: 1.1 (creativity level)
• Density: 0.7 (musical complexity)
• Brightness: 0.6-0.8 (tonal quality)
• Guidance: 4.0 (prompt adherence)
• Mode: QUALITY (vs DIVERSITY)

Controls

While music is playing:

• Ctrl+C: Stop music and exit
• Music streams continuously with real-time generation
• Adaptive to coding rhythm and flow

Requirements

• Node.js 16+
• Valid Gemini API key with Lyria RealTime access
• Internet connection for real-time streaming

Notes

• Experimental Model: Lyria RealTime is currently an experimental model
• Instrumental Only: Generates instrumental music only
• Watermarked: All output follows Google's Responsible AI principles
• Real-time: Music generation adapts and continues based on your prompts

Troubleshooting

Connection Issues:

• Verify your Gemini API key is valid
• Check internet connection
• Ensure Lyria RealTime model access

Audio Issues:

• The CLI simulates audio streaming (actual implementation requires proper WebSocket setup)
• For real audio playback, additional system audio libraries may be needed

Development

This project demonstrates the structure for integrating with Gemini Lyria RealTime. The actual WebSocket implementation requires:

1. Proper authentication with Gemini API
2. WebSocket connection to the Lyria RealTime endpoint
3. Audio buffer management for smooth playback
4. Real-time prompt and configuration updates

License

MIT

Happy Coding with AI Music! 🎵