home / mcp / deepseek mcp server

Deepseek MCP Server

A Model Context Protocol (MCP) server implementation connecting Claude Desktop with DeepSeek's language models (R1/V3)

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "66julienmartin-mcp-server-deepseek_r1": {
      "command": "node",
      "args": [
        "/path/to/deepseek-r1-mcp/build/index.js"
      ],
      "env": {
        "DEEPSEEK_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

You run a Deepseek R1 MCP Server to enable clients to generate contextually aware responses using the Deepseek R1 model. This server integrates with MCP clients, supports configurable generation parameters, and provides reliable error handling and Claude Desktop compatibility for streamlined workflows.

How to use

You connect to the server through an MCP client by selecting the local stdio server option and pointing the client to the running process. Start the server, then send prompts and receive responses via the MCP protocol. You can adjust generation parameters such as the maximum number of tokens and temperature to control creativity and length. The server is designed to handle errors gracefully and provide clear feedback to the client if something goes wrong. If you use Claude Desktop, you can configure it to load this MCP server as a local stdio service for convenient access.

How to install

Prerequisites you need before installation:

Node.js (v18 or higher)
npm
Claude Desktop
Deepseek API key

Step-by-step installation flow you should follow exactly as shown here to ensure the server runs correctly.

# Clone the MCP server package
git clone https://github.com/66julienmartin/MCP-server-Deepseek_R1.git
cd deepseek-r1-mcp

# Install dependencies
npm install

# Copy environment template and add your API key
cp .env.example .env  # Then add your API key

# Build the project for production
npm run build

Additional setup for local operation

{
  "mcpServers": {
    "deepseek_r1": {
      "command": "node",
      "args": ["/path/to/deepseek-r1-mcp/build/index.js"],
      "env": {
        "DEEPSEEK_API_KEY": "your-api-key"
      }
    }
  }
}

Available tools

textGeneration

Advanced text generation with Deepseek R1 using an up-to-8192 token context window and configurable parameters.

modelSwitch

Switch between Deepseek R1 and Deepseek V3 by modifying the model name in the source configuration.