GLPI MCP Server

A Python Model Context Protocol (MCP) server for GLPI IT Service Management, using SSE transport so it can be hosted as a persistent HTTP service and connected to by any MCP client (Claude Desktop, Claude Code, etc.).

Features

Full ITIL coverage across 80+ tools:

| Category | Tools | |---|---| | Tickets | list, get, create, update, delete, add followup/task/solution, assign | | Problems | list, get, create, update | | Changes | list, get, create, update | | Computers | list, get, create, update, delete | | Software | list, get, create | | Network Equipment | list, get | | Printers / Monitors / Phones | list, get | | Knowledge Base | list, get, search, create | | Contracts | list, get, create | | Suppliers | list, get, create | | Locations | list, get, create | | Projects | list, get, create, update | | Users | list, get, search, create | | Groups | list, get, create, add user | | Categories / Entities / Documents | list, get | | Statistics | ticket stats by status, asset inventory counts | | Session | get full session info | | Search | advanced search by criteria |

Quick start

Environment variables

| Variable | Required | Description | |---|---|---| | GLPI_URL | Yes | Base URL of your GLPI instance, e.g. https://glpi.example.com | | GLPI_APP_TOKEN | Recommended | Application token (configured in GLPI Setup → API) | | GLPI_USER_TOKEN | One of these | Personal API token for the user | | GLPI_USERNAME | One of these | Username (used with GLPI_PASSWORD) | | GLPI_PASSWORD | One of these | Password | | HOST | No | Bind host (default 0.0.0.0) | | PORT | No | Bind port (default 8000) |

Run with Docker (recommended)

docker run -d \
  -p 8000:8000 \
  -e GLPI_URL=https://glpi.example.com \
  -e GLPI_APP_TOKEN=your_app_token \
  -e GLPI_USER_TOKEN=your_user_token \
  manikumar1/glpi-mcp:latest

Run locally

pip install -r requirements.txt

export GLPI_URL=https://glpi.example.com
export GLPI_APP_TOKEN=your_app_token
export GLPI_USER_TOKEN=your_user_token

python server.py

The server starts at http://localhost:8000/sse.

Docker Compose

services:
  glpi-mcp:
    image: manikumar1/glpi-mcp:latest
    ports:
      - "8000:8000"
    environment:
      GLPI_URL: https://glpi.example.com
      GLPI_APP_TOKEN: your_app_token
      GLPI_USER_TOKEN: your_user_token
    restart: unless-stopped

Connect to Claude

Claude Desktop (claude_desktop_config.json)

{
  "mcpServers": {
    "glpi": {
      "url": "http://localhost:8000/sse"
    }
  }
}

Claude Code (CLI)

claude mcp add glpi --transport sse http://localhost:8000/sse

GLPI API setup

1. In GLPI, go to Setup → General → API
2. Enable the REST API
3. Create an Application Token for this server
4. Either use a User Token (from user preferences) or provide GLPI_USERNAME + GLPI_PASSWORD

Build

docker build -t glpi-mcp .

Docker image

Pre-built images are published to Docker Hub on every push to main and on version tags:

manikumar1/glpi-mcp:latest
manikumar1/glpi-mcp:main
manikumar1/glpi-mcp:v1.0.0

License

MIT