hidennsim 1.1.3

MCP server with JAX-based numerical tools

HIDENNSIM - MCP Server with JAX Integration

A Model Context Protocol (MCP) server providing JAX-based numerical computation tools.

Features

• MCP server with JAX-powered numerical tools (add_jax, subtract_jax, multiply_jax)
• Cython binary compilation for optimized performance
• Cross-platform support (Windows, macOS, Linux)
• Available on PyPI - simple pip install installation
• Integration guides for Claude Desktop and Cursor IDE

Quick Start

Installation Checklist

Follow these steps in order for successful setup:

• Step 1: Create virtual environment (python -m venv hidennsim-env)
• Step 2: Activate virtual environment
• Step 3: Install hidennsim (pip install hidennsim)
• Step 4: Install JAX (pip install "jax[cpu]")
• Step 5: Verify JAX works (python -c "import jax")
• Step 6: Configure your LLM (Claude Desktop or Cursor IDE) with full path to hidennsim.exe
• Step 7: Restart your LLM application

1. Create Virtual Environment (Recommended)

Creating a virtual environment isolates HIDENNSIM and its dependencies from your system Python.

macOS/Linux:

python3 -m venv hidennsim-env
source hidennsim-env/bin/activate

Windows (PowerShell):

python -m venv hidennsim-env
.\hidennsim-env\Scripts\Activate.ps1

Windows (Command Prompt):

python -m venv hidennsim-env
hidennsim-env\Scripts\activate.bat

Note: Your terminal prompt should now show (hidennsim-env) indicating the virtual environment is active.

2. Install HIDENNSIM

Important: Ensure your virtual environment is activated before running pip install commands.

pip install hidennsim

3. Install JAX (REQUIRED)

CRITICAL: JAX must be installed in the same virtual environment as HIDENNSIM. The MCP server will fail to start without it.

CPU-Only (Recommended for most users):

pip install "jax[cpu]"

GPU-Accelerated (Linux or WSL2 only):

WINDOWS USERS: JAX GPU (CUDA) support does NOT work on native Windows. GPU acceleration requires either:

• Linux (native installation)
• WSL2 (Windows Subsystem for Linux 2)

For Linux or WSL2 with NVIDIA GPU and CUDA 12:

pip install "jax[cuda12]"

Verify Installation:

python -c "import jax; print('JAX', jax.__version__, 'installed successfully')"

4. Configure LLM

Claude Desktop Configuration

1. Open Settings: Click Claude icon → Settings → Developer tab → Edit Config

2. Add MCP Server Configuration:

   macOS/Linux:

   {
     "mcpServers": {
       "hidennsim": {
         "command": "/full/path/to/hidennsim-env/bin/hidennsim",
         "args": []
       }
     }
   }
   

   Windows:

   {
     "mcpServers": {
       "hidennsim": {
         "command": "C:\\full\\path\\to\\hidennsim-env\\Scripts\\hidennsim.exe",
         "args": []
       }
     }
   }
   

   Important: Replace the path with the actual absolute path to your virtual environment.

3. Save and Restart Claude Desktop

Cursor IDE Configuration

1. Press Ctrl+Shift+P (Windows/Linux) or Cmd+Shift+P (macOS)
2. Type "Settings" and select Preferences: Open Settings (UI)
3. Search for "MCP" and add the same configuration as above
4. Save and restart Cursor IDE

Configuration File Locations

Claude Desktop:

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

Cursor IDE:

• macOS: ~/Library/Application Support/Cursor/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json
• Windows: %APPDATA%\Cursor\User\globalStorage\saoudrizwan.claude-dev\settings\cline_mcp_settings.json
• Linux: ~/.config/Cursor/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json

Available Tools

Tool	Description	Parameters
add_jax	Floating-point addition using JAX	a, b (numbers)
subtract_jax	Floating-point subtraction (a - b)	a, b (numbers)
multiply_jax	Floating-point multiplication (a * b)	a, b (numbers)

Troubleshooting

MCP Server Fails to Start

Error: "ModuleNotFoundError: No module named 'jax'"

JAX is not installed. Fix:

# Activate virtual environment first, then:
pip install "jax[cpu]"

Error: "spawn ... hidennsim.exe ENOENT"

The executable path is incorrect. Verify:

1. Check the path exists:
   • Windows: C:\path\to\hidennsim-env\Scripts\hidennsim.exe
   • macOS/Linux: /path/to/hidennsim-env/bin/hidennsim
2. Use absolute paths (not relative) in your config
3. Restart your LLM application after config changes

Verify Installation

# Check hidennsim
pip show hidennsim

# Check JAX
pip show jax

# Test MCP server
hidennsim --help

System Requirements

• Python: 3.10 or higher
• Operating System: Windows 10+, macOS 12+, or Linux (Ubuntu 20.04+)
• RAM: 4GB minimum (8GB recommended for GPU acceleration)

Documentation

• Installation Guide
• Configuration Guide
• Troubleshooting

License

MIT License - See LICENSE file for details.

Support

• Issues: https://github.com/yourusername/hidennsim/issues