CLIO Kit — MCP Servers

16 specialized MCP servers for scientific computing — 150+ tools

CLIO Kit provides AI agents with direct access to HPC infrastructure, scientific data formats, and research workflows through Model Context Protocol (MCP) servers.

tip

For the interactive MCP showcase, visit docs.iowarp.ai.

Quick Start

# Install and run any MCP server
uvx clio-kit mcp-server hdf5
uvx clio-kit mcp-server pandas
uvx clio-kit mcp-server slurm

# List all available servers
uvx clio-kit mcp-servers

Available Servers

Data I/O & Visualization (5 servers, 56+ tools)

Server	Tools	Description
HDF5	27	HDF5 file operations with AI-powered insights, LRU caching, streaming
Pandas	15	CSV data analysis, statistical summaries, data cleaning
ADIOS	5	Scientific data I/O for BP5 format
Parquet	4	High-performance columnar data processing
ParaView	29	3D scientific visualization, isosurfaces, streamlines

HPC Integration (4 servers, 59+ tools)

Server	Tools	Description
Slurm	13	Job submission, monitoring, cluster analytics
Jarvis	25+	HPC workflow orchestration, pipeline management
Lmod	10	HPC module system management
Node Hardware	11	CPU, GPU, memory, disk, network monitoring

Research Discovery (2 servers, 16+ tools)

Server	Tools	Description
ArXiv	13	Paper search, metadata, BibTeX export, citation networks
NDP	3	National Data Platform dataset discovery

System Monitoring (2 servers, 14+ tools)

Server	Tools	Description
Darshan	10	I/O performance analysis, bandwidth, bottleneck detection
ChronoLog	4	Session logging and context sharing

Utilities (3 servers, 20+ tools)

Server	Tools	Description
Plot	6	Data visualization (line, bar, scatter, heatmap)
Compression	1	File compression utilities
Parallel Sort	13	Log processing, pattern detection, filtering

IDE Integration

Cursor

Add to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "hdf5-mcp": {
      "command": "uvx",
      "args": ["clio-kit", "mcp-server", "hdf5"]
    },
    "pandas-mcp": {
      "command": "uvx",
      "args": ["clio-kit", "mcp-server", "pandas"]
    },
    "slurm-mcp": {
      "command": "uvx",
      "args": ["clio-kit", "mcp-server", "slurm"]
    }
  }
}

Cursor MCP docs →

Claude Code

claude mcp add hdf5-mcp -- uvx clio-kit mcp-server hdf5
claude mcp add pandas-mcp -- uvx clio-kit mcp-server pandas
claude mcp add slurm-mcp -- uvx clio-kit mcp-server slurm

Claude Code MCP docs →

VS Code

Add to your VS Code MCP config:

"mcp": {
  "servers": {
    "hdf5-mcp": {
      "type": "stdio",
      "command": "uvx",
      "args": ["clio-kit", "mcp-server", "hdf5"]
    },
    "pandas-mcp": {
      "type": "stdio",
      "command": "uvx",
      "args": ["clio-kit", "mcp-server", "pandas"]
    }
  }
}

VS Code MCP docs →

Claude Desktop

Edit claude_desktop_config.json:

{
  "mcpServers": {
    "hdf5-mcp": {
      "command": "uvx",
      "args": ["clio-kit", "mcp-server", "hdf5"]
    },
    "arxiv-mcp": {
      "command": "uvx",
      "args": ["clio-kit", "mcp-server", "arxiv"]
    }
  }
}

Claude Desktop MCP docs →

Repository

• GitHub: github.com/iowarp/clio-kit
• PyPI: pypi.org/project/clio-kit
• Interactive Showcase: docs.iowarp.ai