mloop-mcp

MCP (Model Context Protocol) server for MLoop CLI.

Enables AI clients (Claude, Cursor, ironhive-cli) to perform MLOps tasks using MLoop's AutoML capabilities.

Installation

``bash
npm install -g @iyulab/mloop-mcp
`

Or run directly with npx:

`bash
npx @iyulab/mloop-mcp
`

Prerequisites

- Node.js 18+
- MLoop CLI installed and available in PATH
`bash
dotnet tool install -g mloop
`

$3

If MLoop is not in PATH, set the MLOOP_PATH environment variable:

`bash

Windows

set MLOOP_PATH=D:\lib\mloop.exe

Linux/macOS

export MLOOP_PATH=/usr/local/bin/mloop
`

Configuration

$3

Add to your claude_desktop_config.json:

`json
{
"mcpServers": {
"mloop": {
"command": "npx",
"args": ["@iyulab/mloop-mcp"]
}
}
}
`

$3

Add to .ironhive/plugins.yaml:

`yaml
plugins:
mloop:
transport: stdio
command: npx
args:
- "@iyulab/mloop-mcp"
`

Available Tools

| Tool | Description |
|------|-------------|
| mloop_train | Train ML models using AutoML |
| mloop_predict | Run predictions with trained models |
| mloop_list | List experiments and their metrics |
| mloop_promote | Promote an experiment to production |
| mloop_info | Analyze and profile datasets |
| mloop_status | Show project status |
| mloop_compare | Compare multiple experiments |
| mloop_evaluate | Evaluate model performance |
| mloop_serve | Start REST API server |

Tool Examples

$3

`json
{
"tool": "mloop_train",
"arguments": {
"projectPath": "/path/to/project",
"dataFile": "datasets/train.csv",
"label": "target",
"task": "binary-classification",
"time": 60
}
}
`

$3

`json
{
"tool": "mloop_predict",
"arguments": {
"projectPath": "/path/to/project",
"dataFile": "datasets/test.csv",
"output": "predictions/output.csv"
}
}
`

$3

`json
{
"tool": "mloop_list",
"arguments": {
"projectPath": "/path/to/project",
"showAll": true
}
}
`

$3

`json
{
"tool": "mloop_promote",
"arguments": {
"projectPath": "/path/to/project",
"experimentId": "exp-003",
"force": true
}
}
`

$3

`json
{
"tool": "mloop_info",
"arguments": {
"dataFile": "datasets/train.csv",
"projectPath": "/path/to/project"
}
}
`

Development

`bash

Install dependencies

npm install

Build

npm run build

Run locally

npm start

Test with MCP Inspector

npx @modelcontextprotocol/inspector node build/index.js
`

Architecture

`
ironhive-cli ──[MCP/STDIO]──> mloop-mcp ──[subprocess]──> mloop CLI ──> ML.NET
`

- Pure Bridge Pattern: 1:1 CLI mapping, no business logic, stateless
- STDIO Transport: Simple process communication
- Subprocess Execution: Each tool call spawns mloop` CLI

License

MIT