PLUMED2 MCP Server / PLUMED2 MCP 服务器

by PhelanShao

209 downloads
Not rated
GitHub

About

PLUMED2 Smart Input Generator - An MCP-based assistant for molecular dynamics enhanced sampling configurations that automatically generates optimized PLUMED input files, supporting metadynamics, umbrella sampling, and other enhanced sampling methods with built-in syntax validatio

Details

Author
PhelanShao
Downloads
209
Categories
Other

- Generate complete PLUMED input files with collective variables and bias methods
- Access pre-defined templates for common simulation scenarios (protein folding, drug binding, phase transitions, nucleation)
- Get intelligent performance optimization suggestions for large systems
- Validate PLUMED syntax with detailed error reporting and reference checking
- Fully English codebase with professional documentation

Setting up with Highlight

This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:

  1. Download and install Highlight from highlightai.com/download
  2. Navigate to the plugins tab and select "Add Custom Plugin"
  3. Configure the plugin with the settings below
    Plugin Name PLUMED2 MCP Server / PLUMED2 MCP 服务器
    Command (node, npx, python, etc.)

    Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.

  4. Enable "Start Automatically" if you want the plugin to start when Highlight launches

From the repository

Install Python 3.11 or higher, install dependencies via pip install -r requirements.txt, then add a JSON configuration to your MCP client (e.g., Claude Desktop, Cline) specifying the command python -m plumed_mcp_server.main. The server exposes four tools: generate_plumed_input, validate_plumed_syntax, optimize_plumed_performance, and get_plumed_template. AI assistants invoke these tools based on natural language prompts.

Claude Desktop / Cursor

Paste into your MCP client config file to install this server.

{
    "mcpServers": {
        "plumed2 mcp server / plumed2 mcp \u670d\u52a1\u5668": {
            "plumed2-mcp-server": {
                "command": "python",
                "args": [
                    "-m",
                    "pytest",
                    "tests/",
                    "-v"
                ]
            }
        }
    }
}

McpServers

{
    "plumed2-mcp-server": {
        "command": "python",
        "args": [
            "-m",
            "pytest",
            "tests/",
            "-v"
        ]
    }
}

PLUMED2 MCP Server / PLUMED2 MCP 服务器

English | 中文

---

English

Overview

The PLUMED2 MCP Server is a Model Context Protocol (MCP) server that provides AI assistants with powerful tools for generating, validating, and optimizing PLUMED input files for molecular dynamics simulations. This server enables seamless integration of PLUMED expertise into AI workflows.

Features

- 🔧 Input File Generation: Generate complete PLUMED input files with collective variables, bias methods, and output configurations
- 📋 Template System: Access pre-defined templates for common simulation scenarios (protein folding, drug binding, phase transitions, nucleation)
- ⚡ Performance Optimization: Get intelligent suggestions to optimize PLUMED performance for large systems
- ✅ Syntax Validation: Validate PLUMED syntax with detailed error reporting and reference checking
- 🌐 International: Fully English codebase with professional documentation

Installation

Prerequisites

- Python 3.11 or higher
- pip package manager

Install Dependencies

cd plumed_mcp_server
pip install -r requirements.txt

MCP Client Configuration

Add the following configuration to your MCP client settings (e.g., Claude Desktop, Cline, etc.):

{
  "mcpServers": {
    "plumed2": {
      "command": "python",
      "args": [
        "-m", "plumed_mcp_server.main"
      ],
      "cwd": "/path/to/plumed_mcp_server",
      "alwaysAllow": [
        "generate_plumed_input",
        "validate_plumed_syntax", 
        "optimize_plumed_performance",
        "get_plumed_template"
      ]
    }
  }
}

Note: Replace /path/to/plumed_mcp_server with the actual path to your installation directory.

Available Tools

1. generate_plumed_input

Generate complete PLUMED input files based on simulation requirements.

Parameters:
- collective_variables (required): List of collective variables to compute
- bias_methods (optional): List of bias methods to apply
- output_config (optional): Output configuration settings
- system_info (optional): System information for optimization
- molinfo_config (optional): Molecular information settings
- group_definitions (optional): Custom group definitions

2. validate_plumed_syntax

Validate PLUMED input syntax and check for common errors.

Parameters:
- plumed_input (required): PLUMED input content to validate
- number_of_replicas (optional): Number of replicas for @replicas syntax validation

3. optimize_plumed_performance

Get performance optimization suggestions for PLUMED simulations.

Parameters:
- plumed_input (required): PLUMED input content to analyze
- system_info (optional): System information (number of atoms, simulation length, etc.)

4. get_plumed_template

Retrieve pre-defined or custom PLUMED templates.

Parameters:
- template_name (required): Name of the template to retrieve
- system_type (optional): Type of system for template customization

Usage Examples

Example 1: Generate a Simple Distance CV with Metadynamics

Prompt to AI:

Generate a PLUMED input file for metadynamics simulation of a protein. I want to use the distance between atoms 1 and 2 as a collective variable, with metadynamics bias. Include output every 100 steps.

Expected AI Response:
The AI will use the generate_plumed_input tool with parameters like:

{
"collective_variables": [
{
"type": "DISTANCE",
"label": "d1",
"atoms": [1, 2]
}
],
"bias_methods": [
{
"type": "METAD",
"arg": "d1",
"sigma": 0.1,
"height": 1.0,
"pace": 500
}
],
"output_config": {
"stride": 100,
"file": "colvar.dat"
}
}

Generated Output:
```

No reviews yet — be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.