SonicBridge MCP: Model Context Protocol Music Aesthetic Server

English | 简体中文

Note

sonic-bridge-mcp is a lightweight, ultra-fast, pure-Rust Model Context Protocol (MCP) server that empowers AI companion agents with "physical listening double-ears (超级听感)" under the LRMD protocol. It provides direct tools to analyze and compare musical versions with millisecond-level precision.

---

🚀 Quick One-Click Installation

If you are using a standard Linux/macOS environment with cargo and bash in your PATH, you can configure everything in 3 seconds using our one-click installer:

# Clone the repository (if not already local)
git clone https://github.com/Xuepoo/sonic-bridge-mcp.git
cd sonic-bridge-mcp

# Run the automated installer
./install.sh

What install.sh does:

1. Compiles the Rust MCP server in optimized release mode.
2. Installs the executable binary directly into your XDG-compliant user binary directory: ~/.local/bin/sonic-bridge-mcp.
3. Creates and automatically configures your Claude Desktop config file (~/.config/Claude/claude_desktop_config.json) with zero absolute-path dependencies.

---

🛠️ Manual Installation & Configurations

If you prefer to configure manually or are on a different environment, follow the steps below:

1. Build the Binary

cargo build --release

The optimized executable will be generated at target/release/sonic-bridge-mcp. You can copy this executable to any directory in your system $PATH (e.g., ~/.local/bin/ or /usr/local/bin/).

2. Configure Claude Desktop

Modify (or create) your Claude Desktop configuration file:

• Linux/macOS: ~/.config/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json

Append the following server config:

{
  "mcpServers": {
    "sonic-bridge": {
      "command": "sonic-bridge-mcp",
      "args": []
    }
  }
}

Note: If sonic-bridge-mcp is not in your global system PATH, replace "command" with the absolute path to your compiled binary.

3. Configure Cursor / Windsurf / Copilot IDE

In your Cursor settings (Settings -> Features -> MCP), click + Add New MCP Server:

• Name: sonic-bridge
• Type: command (Stdio)
• Command: sonic-bridge-mcp
• Args: (Leave Blank)

---

🧠 Provided AI Tools Reference

Once connected, your AI Agent will automatically discover and call the following acoustic tools:

1. analyze_music

Analyze a local audio track (.mp3, .flac, .wav, etc.) and return an LLM-Readable Music Descriptor (LRMD) report outlining temporal structures, chords, dynamics, and timbres.

• Arguments:
  • filepath (string, required): The absolute path to the target music file.
  • onset_mode (boolean, optional): Set to true to use dynamic transient-driven onset segmentation (recommended for complex electronic/pop music). false defaults to static intervals.
• Return: A Markdown formatted semantic music description.

2. compare_music

Compare two different versions of the same track (e.g., a studio master vs. an acoustic cover version) using Dynamic Time Warping (DTW) to align temporal indices and contrast performance affects.

• Arguments:
  • file_a (string, required): The absolute path to the original track.
  • file_b (string, required): The absolute path to the comparative cover track.
• Return: A detailed comparative warp-matrix table tracking structural difference.

---

🌊 Under the Hood: Pure DSP Aesthetics

• Token-Optimized Chunk Merger: Automatic adjacent slice fusion that compresses redundant micro-segments into Phrase Blocks, cutting down downstream Agent token consumption by over 90%.
• Crest Factor Adaptive Classification: Combines wave peak-to-RMS ratios with sliding relative RMS levels to detect the "breathing" and dynamic fluctuations of modern brickwall-limited masters.
• 200Hz - 2kHz Bandpass Chroma Projection: Restricts harmonic tracking to the human vocal and melodic mid-range, preventing heavy sub-bass kick drums from muddying minor scale classifications.

---

📄 License

This project is licensed under the MIT License.