Configuration

This guide explains how to configure the Math-Physics-ML MCP servers for use with Claude Desktop or Claude Code.

Available MCP Servers

The system provides five specialized MCP servers:

1. math-mcp - Symbolic algebra and GPU-accelerated numerical computing
2. quantum-mcp - Wave mechanics and Schrödinger equation simulations
3. molecular-mcp - Classical molecular dynamics simulations
4. neural-mcp - Neural network training and experimentation
5. llm-mcp - Language model training, fine-tuning, and experimentation

Prerequisites

Before configuring the MCP servers, ensure they are installed:

# Verify installation
ls -la .venv/bin/*-mcp

# You should see:
# - .venv/bin/math-mcp
# - .venv/bin/quantum-mcp
# - .venv/bin/molecular-mcp
# - .venv/bin/neural-mcp
# - .venv/bin/llm-mcp

If the servers are not installed, run:

uv sync --all-extras
uv pip install --python .venv/bin/python3 -e shared/mcp-common -e shared/compute-core
uv pip install --python .venv/bin/python3 -e servers/math-mcp -e servers/quantum-mcp
uv pip install --python .venv/bin/python3 -e servers/molecular-mcp -e servers/neural-mcp -e servers/llm-mcp

Configuration Methods

Claude Desktop Configuration

Claude Desktop is the recommended method for GUI-based usage.

Configuration File Location:

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

Configuration:

Add the following to your claude_desktop_config.json file:

{
  "mcpServers": {
    "math-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/home/beagle/work/math-mcp",
        "math-mcp"
      ]
    },
    "quantum-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/home/beagle/work/math-mcp",
        "quantum-mcp"
      ]
    },
    "molecular-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/home/beagle/work/math-mcp",
        "molecular-mcp"
      ]
    },
    "neural-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/home/beagle/work/math-mcp",
        "neural-mcp"
      ]
    },
    "llm-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/home/beagle/work/math-mcp",
        "llm-mcp"
      ]
    }
  }
}

Important: Replace /home/beagle/work/math-mcp with the absolute path to your math-mcp installation directory.

Verification:

1. Restart Claude Desktop
2. Start a new conversation
3. The MCP servers should appear in the tools panel
4. Try using a tool: "Create a 5x5 matrix of random numbers" (uses math-mcp)

Claude Code CLI Configuration

For command-line usage with Claude Code:

# Add servers manually using Claude Code CLI
claude-code config add-mcp-server math-mcp "uv run --directory $(pwd) math-mcp"
claude-code config add-mcp-server quantum-mcp "uv run --directory $(pwd) quantum-mcp"
claude-code config add-mcp-server molecular-mcp "uv run --directory $(pwd) molecular-mcp"
claude-code config add-mcp-server neural-mcp "uv run --directory $(pwd) neural-mcp"
claude-code config add-mcp-server llm-mcp "uv run --directory $(pwd) llm-mcp"

Alternatively, use the provided mcp-servers.json configuration file in the project root.

Verification:

# List available MCPs
claude-code mcp list

# Check server status
claude-code mcp status

Direct Execution (Testing)

For testing and development, you can run any MCP server directly:

# Activate the virtual environment
source .venv/bin/activate  # Linux/macOS
# or
.venv\Scripts\activate     # Windows

# Run any server directly
math-mcp
quantum-mcp
molecular-mcp
neural-mcp
llm-mcp

# Or with uv (no activation needed)
uv run math-mcp
uv run quantum-mcp
uv run molecular-mcp
uv run neural-mcp
uv run llm-mcp

Tool Discovery

Each MCP server provides an info tool for progressive discovery of capabilities:

Math MCP

info(topic="overview")           # See all categories
info(topic="symbolic")           # Symbolic algebra tools
info(topic="numerical")          # Numerical computing tools
info(topic="transforms")         # FFT and transforms
info(topic="optimization")       # Optimization tools
info(topic="matrix_multiply")    # Specific tool help

Quantum MCP

info(topic="overview")           # All quantum tools
info(topic="solve_schrodinger") # Schrödinger solver help

Molecular MCP

info()                          # All molecular dynamics tools

Neural MCP

info()                          # All neural network tools

LLM MCP

info(topic="overview")           # All LLM tools
info(topic="models")             # Model creation tools
info(topic="tokenizers")         # Tokenization tools
info(topic="datasets")           # Dataset tools
info(topic="training")           # Training tools

GPU Support

All MCP servers support GPU acceleration with automatic CPU fallback:

• GPU Available: Set use_gpu=True in tool calls for 10-100x speedup
• No GPU: Automatically falls back to CPU (NumPy instead of CuPy)
• Check Status: GPU Manager logs CUDA availability at startup

To verify GPU availability:

# Check CUDA availability
python -c "import torch; print(torch.cuda.is_available())"
python -c "import cupy; print(cupy.cuda.is_available())"

Configuration Files

The system uses the following configuration files:

• mcp-servers.json - Pre-configured server definitions
• config.kdl - GPU memory settings and runtime configuration
• pyproject.toml - Package dependencies and entry points

GPU Memory Configuration

Edit config.kdl to customize GPU memory settings:

gpu {
    memory_pool_size "500MB"
    memory_growth true
}

Troubleshooting

Server Won't Start

# Check Python version (requires 3.11+)
python --version

# Reinstall dependencies
uv sync

# Reinstall servers
uv pip install -e servers/math-mcp -e servers/quantum-mcp \
               -e servers/molecular-mcp -e servers/neural-mcp

Entry Points Not Found

# Verify installation
pip show math-mcp quantum-mcp molecular-mcp neural-mcp

# Check entry points
which math-mcp quantum-mcp molecular-mcp neural-mcp

# Should show paths like:
# /path/to/math-mcp/.venv/bin/math-mcp

GPU Not Detected

# Check CUDA availability
python -c "import torch; print(torch.cuda.is_available())"
python -c "import cupy; print(cupy.cuda.is_available())"

# If False, install CUDA toolkit or use CPU mode (automatic fallback)

Tools Not Appearing in Claude Desktop

1. Ensure the configuration file path is correct
2. Verify the absolute path to math-mcp directory is correct
3. Restart Claude Desktop completely
4. Check Claude Desktop logs for error messages

Next Steps

Once configured, proceed to the Quick Start guide to learn how to use the MCP tools.