LX Connect MCP Server

What is LX Connect MCP Server?

LX Connect MCP Server is a free, hosted MCP server from Co-Orga GmbH that gives AI assistants knowledge of the LX Connect interface for Lexware: its OData read model, data warehouse tables, field descriptions and REST API documentation. It needs no sign-in and has no access to your own data.

LX Connect itself is software installed on a Lexware server. It exposes Lexware Pro and Premium (on-premises) data as a SQL data warehouse and as a web API that can read and also write records such as customers, orders and invoices. It is aimed at system integrators, developers and IT departments that connect Lexware to other systems. The MCP server is the "knowledge" side of that product: it helps an assistant write correct queries and integration code, which you then run against your own installation.

How to use LX Connect MCP Server?

There is nothing to install and no API key. The server uses Streamable HTTP at:

https://mcp.lxconnect.de/

Co-Orga describes the connection as needing "no key, no header, no approval". Any client that speaks MCP over HTTP can use the URL as is.

Claude Code

claude mcp add --transport http lx-connect https://mcp.lxconnect.de/

Run /mcp inside Claude Code to confirm the server is connected.

Claude Desktop and claude.ai

According to the vendor, LX Connect is listed in Claude's connector directory: open Settings, then Connectors, choose Browse connectors, select LX Connect and connect. You can also add it manually as a custom connector with the URL above; no authentication is needed.

Cursor

Add the server to ~/.cursor/mcp.json or the project's .cursor/mcp.json:

{
  "mcpServers": {
    "lx-connect": {
      "url": "https://mcp.lxconnect.de/"
    }
  }
}

VS Code

Add it to .vscode/mcp.json:

{
  "servers": {
    "lx-connect": {
      "type": "http",
      "url": "https://mcp.lxconnect.de/"
    }
  }
}

ChatGPT

Co-Orga notes that custom connections in ChatGPT need a Business or Enterprise account: enable developer mode, add a new app with the server URL and choose "No Auth".

Key features of LX Connect MCP Server

• Free to use, with no registration, API key or OAuth flow.
• Describes 81 data areas of the LX Connect OData read model with their real field names.
• Documents the LX Connect data warehouse: 82 tables with descriptive names and field-level business descriptions, grouped by module (Warenwirtschaft, Buchhaltung, Anlagenverwaltung, Allgemein).
• Covers the 31 write endpoints of the REST API with required fields and example calls.
• Includes OData guidance such as aggregation, paging and worked query patterns.
• Answers product questions: features, system requirements and pricing.
• No data access: the server has no database connection, session or identity, stores no requests and makes no outbound calls.

LX Connect MCP Server tools

The server returns seven tools, all with German names and descriptions.

OData read model

• entitaeten_auflisten lists every entity set of the LX Connect read model, grouped by module and topic, in one call.
• schema_beschreiben returns the fields (name, type, optional, key) and navigation properties of one entity set; the navigations are exactly the targets allowed in $expand.

Data warehouse (SQL target database)

• lxdw_tabellen_auflisten lists all tables LX Connect creates in the customer's target database, with their descriptions. This is the view used for Power BI and custom SQL reports.
• lxdw_tabelle_beschreiben returns the columns of one table (SQL type, nullability, primary key, description) plus its foreign keys and referencing tables.
• lxdw_feld_suchen searches column names and descriptions across all tables, which answers questions like "which table holds the IBAN or the cost center?".

Documentation

• doku_suchen searches product information, the service's help chapters and the REST endpoint list; without a search term it returns the table of contents.
• doku_lesen returns the full text of one document found with doku_suchen.

Use cases of LX Connect MCP Server

• Writing OData queries against your own LX Connect installation: "Which fields and navigations does the customer entity set have, and how do I expand its addresses?"
• Building Power BI or SQL reports: "In which data warehouse table is the delivery date stored?"
• Preparing write calls to the REST API: "What are the required fields to create an order through LX Connect?"
• Checking project fit before buying: "What are the system requirements for LX Connect and what does it cost?"

Troubleshooting LX Connect MCP Server

• The assistant cannot see my data. This is by design. The server only describes the interface; queries and write calls are templates you run in your own code against your LX Connect installation.
• Answers mix up OData and SQL names. Ask explicitly for the OData read model (entitaeten_auflisten, schema_beschreiben) or the data warehouse (lxdw_* tools).
• ChatGPT will not let me add it. Custom connections there require a Business or Enterprise account with developer mode enabled.