Matrix42 MCP Server

The Matrix42 MCP Server connects AI assistants with a Matrix42 instance through the Model Context Protocol. The assistant can find the right web service, read the data model, query records and search the service desk. It only reads unless writing is switched on and every write is shown as a preview first.

npx -y matrix42-mcp
Matrix42 MCP Server

Why an assistant needs it

A typical Matrix42 instance has about 190 web services with some 1,100 operations, a data model of around 800 data definitions and 240 configuration items. An AI assistant cannot know which of them exist in your instance, so it guesses URLs, attribute names and filters. The server gives it a way to look things up instead: it finds the right web service, reads the real data model and checks every column against your schema before a query runs.

What it can do

  • Discover the API: all operations with their request and response contracts
  • Understand the data model: data definitions, configuration items, pickup values and relations
  • Read records: ASQL queries with paging, saved views, journal entries, attachments and links into the Matrix42 web interface
  • Work the service desk: search tickets of any kind, answer service level questions, browse assets, contracts, catalog services and knowledge articles
  • Act on tickets, if you allow it: create, close, take over, forward, pause, reopen, set deadlines and track time

A question like "Which open hardware tickets are past their service level?" gets answered without anyone looking up an ID.

Safe by default

  • The server only reads. Tools that change data are not even offered to the assistant unless you switch writing on.
  • Every write is shown as a preview first and only happens after it is confirmed.
  • The credentials stay in the server. The assistant never sees them.
  • Tickets created through the server get an internal note that says so. Notification emails are off and journal entries are internal by default.

Setting it up

You need Node.js 22.19 or later and a Matrix42 API token, created in the Matrix42 administration for the account the assistant should act as. The server runs over stdio, so your AI client starts it with npx and nothing has to be installed first. For Claude Code:

claude mcp add matrix42 \
  --env M42_HOST=https://matrix42.example.com \
  --env M42_API_TOKEN=your-api-token \
  -- npx -y matrix42-mcp

Claude Desktop, Cursor, VS Code with GitHub Copilot and other MCP clients use the same command with their own configuration file. The README on GitHub has an example for each.

Good to know

The server is listed in the official MCP Registry as io.github.sus-tech-gmbh/matrix42-mcp, where MCP clients and directories find it.

The server is an early release and an independent community project under the MIT licence. It is not affiliated with or supported by Matrix42, so please do not contact Matrix42 support about it. Questions and bug reports go to the issues of the repository on GitHub.

For automation without an assistant, the Matrix42 Node for n8n brings the same Matrix42 instance into n8n workflows.

Examples and the background are in our post Matrix42 MCP Server: Ask Your Service Desk in Plain Language.

Using it in your company?

The code is free. We help with the setup, with changes and with connecting it to your processes.