DeepL MCP Server logo

DeepL MCP Server

by watchdealer-pavel

This is an MCP server that provides DeepL translation capabilities. It allows AI assistants to translate text and list supported languages using the DeepL API.

View on GitHub

Last updated: N/A

DeepL MCP Server

An MCP (Model Context Protocol) server providing DeepL translation capabilities.

Features

This server exposes the following tools via MCP:

  • translate_text: Translates one or more text strings between supported languages using the DeepL API.
  • list_languages: Retrieves the list of languages supported by the DeepL API (either source or target languages).

Prerequisites

  • Node.js and npm/yarn: Required to install dependencies and run the server.
  • DeepL API Key: You need an API key from DeepL. Both Free and Pro plans provide API access. Sign up or learn more at https://www.deepl.com/pro-api.

Installation

  1. Clone the repository:

    git clone https://github.com/watchdealer-pavel/deepl-mcp-server.git
    cd deepl-mcp-server
    
  2. Install dependencies:

    npm install
    # or
    # yarn install
    
  3. Build the server:

    npm run build
    

    This command compiles the TypeScript source code into JavaScript, placing the output in the build/ directory (specifically build/index.js).

Configuration

The server requires your DeepL API key to be provided via the DEEPL_API_KEY environment variable. You need to configure your MCP client (like Cline/Roo Code or the Claude Desktop App) to run this server and pass the environment variable.

Example Configuration:

Below are examples for common MCP clients. Remember to replace /path/to/your/deepl-mcp-server/build/index.js with the actual absolute path to the compiled server file on your system, and YOUR_DEEPL_API_KEY with your real DeepL API key.

Cline / Roo Code (VS Code Extension)

  1. Open your VS Code settings for MCP servers. On macOS, this is typically located at: ~/Library/Application Support/Code/User/globalStorage/rooveterinaryinc.roo-cline/settings/mcp_settings.json (Note: The exact path might vary based on your operating system and VS Code installation type (e.g., Insiders).)

  2. Add the following configuration block under the mcpServers key:

    "deepl-translator": {
      "command": "node",
      "args": ["/path/to/your/deepl-mcp-server/build/index.js"], // <-- IMPORTANT: Replace with the ACTUAL absolute path to build/index.js
      "env": {
        "DEEPL_API_KEY": "YOUR_DEEPL_API_KEY" // <-- IMPORTANT: Replace with your DeepL API Key
      },
      "disabled": false,
      "alwaysAllow": []
    }
    

Claude Desktop App

  1. Open the Claude Desktop App configuration file. On macOS, this is typically located at: ~/Library/Application Support/Claude/claude_desktop_config.json (Note: The exact path might vary based on your operating system.)

  2. Add the following configuration block under the mcpServers key:

    "deepl-translator": {
      "command": "node",
      "args": ["/path/to/your/deepl-mcp-server/build/index.js"], // <-- IMPORTANT: Replace with the ACTUAL absolute path to build/index.js
      "env": {
        "DEEPL_API_KEY": "YOUR_DEEPL_API_KEY" // <-- IMPORTANT: Replace with your DeepL API Key
      },
      "disabled": false,
      "alwaysAllow": []
    }
    

Usage

Once configured, you can invoke the server's tools from your AI assistant using the use_mcp_tool command/tool.

list_languages Example

<use_mcp_tool>
  <server_name>deepl-translator</server_name>
  <tool_name>list_languages</tool_name>
  <arguments>
    {
      "type": "target" // Optional: "source" or "target". Defaults to listing all if omitted.
    }
  </arguments>
</use_mcp_tool>

translate_text Example

<use_mcp_tool>
  <server_name>deepl-translator</server_name>
  <tool_name>translate_text</tool_name>
  <arguments>
    {
      "text": ["Hello world", "How are you?"], // Required: An array of strings to translate
      "target_lang": "DE", // Required: Target language code (e.g., DE, FR, ES)
      "source_lang": "EN" // Optional: Source language code. DeepL will auto-detect if omitted.
    }
  </arguments>
</use_mcp_tool>

License

This project is licensed under the MIT License - see the LICENSE file for details.