Translate

You run this MCP server to translate text using a local MCP endpoint. It supports two transports: stdio for direct integration with Claude Desktop and HTTP for streaming tests in a browser or via curl. This server relies on Lingvanex Translator for the actual translation capability and exposes a simple MCP tool you can call to translate text between languages.

How to use

You can use the server with an MCP client in two ways. In stdio mode, your client connects locally and communicates through the process I/O. In HTTP mode, you run a local HTTP server that accepts requests over HTTP and streams results to the client. Start the server in the mode you prefer, then configure your MCP client to reach the local endpoint.

How to install

Prerequisites: You need Node.js 18 or newer and a package manager like Yarn or npm. You also need a Lingvanex Translator account to obtain an API key for text translation.

Step 1: Install dependencies by running the package manager in the project root.

Step 2: Build the server to produce a runnable artifact.

Step 3: Start the server in your chosen transport mode.

Configuration and running notes

To use stdio transport with Claude Desktop, set the transport mode to stdio and start the server. The server will indicate that the stdio transport is running and that the Translate MCP Server is ready.

To use HTTP transport for testing, set the transport to http and specify a port for the server to listen on. Start the server, then you can ping the endpoint or connect with debugging tools like the MCP Inspector.

Example quick-start sequence in your terminal is shown in the next sections. You can reuse these exact commands to reproduce the setup.

Testing and debugging

Test the basic HTTP health endpoint by querying the local server. You should receive a JSON response indicating an ok status and the transport type.

If you want to inspect the MCP traffic interactively, use the MCP Inspector, connect to the local streamable HTTP URL, and verify the translation tool calls and outputs.

Available tool

translate_text translates a given piece of text from a source language to a target language.

Examples

{
  "tool": "translate_text",
  "args": {
    "text": "Good morning",
    "sourceLang": "en",
    "targetLang": "fr"
  }
}

Available tools

translate_text

Translate text from a source language to a target language. Requires text, sourceLang, and targetLang.