Deepseek Thinker MCP Server

Source

A MCP provider Deepseek reasoning content to MCP-enabled AI Clients, like Claude Desktop. Supports access to Deepseek's CoT from the Deepseek API service or a local Ollama server.

Catalog onlyCatalog onlySTDIO

Overview

Deepseek Thinker MCP is a Model Context Protocol (MCP) provider that delivers Deepseek reasoning content to MCP-enabled AI clients, such as Claude Desktop. It allows access to Deepseek's thought processes via the Deepseek API service or a local Ollama server.

To use Deepseek Thinker MCP, integrate it with an AI client by configuring the claude_desktop_config.json file with the necessary command and environment variables. You can also run it in Ollama mode or configure it for local server use.

Dual Mode Support: OpenAI API mode and Ollama local mode. - Focused Reasoning: Captures and provides reasoning output from Deepseek's thinking process.

Enhancing AI client capabilities with Deepseek's reasoning.
Supporting complex reasoning tasks in AI applications.
Facilitating local AI model interactions through Ollama.

Add to your AI client

Use these steps to connect Deepseek Thinker MCP Server in Cursor, Claude, VS Code, and other MCP-compatible apps. The same JSON appears in the Use with menu above for one-click copy.

Cursor

Add this to your .cursor/mcp.json file in your project root, then restart Cursor.

.cursor/mcp.json

{
  "mcpServers": {
    "deepseek-thinker-mcp-ruixingshi": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-deepseek-thinker-mcp-ruixingshi"
      ]
    }
  }
}

Claude Desktop

Add this server entry to the mcpServers object in your Claude Desktop config, then restart the app.

~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows)

{
  "mcpServers": {
    "deepseek-thinker-mcp-ruixingshi": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-deepseek-thinker-mcp-ruixingshi"
      ]
    }
  }
}

Claude Code

Add this to your project's .mcp.json file. Claude Code will detect it automatically.

.mcp.json (project root)

{
  "mcpServers": {
    "deepseek-thinker-mcp-ruixingshi": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-deepseek-thinker-mcp-ruixingshi"
      ]
    }
  }
}

VS Code (Copilot)

Add this to your .vscode/mcp.json file. Requires the GitHub Copilot extension with MCP support enabled.

.vscode/mcp.json

{
  "servers": {
    "deepseek-thinker-mcp-ruixingshi": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-deepseek-thinker-mcp-ruixingshi"
      ]
    }
  }
}

Windsurf

Add this to your Windsurf MCP config file, then restart Windsurf.

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "deepseek-thinker-mcp-ruixingshi": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-deepseek-thinker-mcp-ruixingshi"
      ]
    }
  }
}

Cline

Open Cline settings, navigate to MCP Servers, and add this server configuration.

Cline MCP Settings (via UI)

{
  "mcpServers": {
    "deepseek-thinker-mcp-ruixingshi": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-deepseek-thinker-mcp-ruixingshi"
      ]
    }
  }
}

FAQ

What should I do if I encounter "MCP error -32001: Request timed out"?

This error indicates that the Deepseek API response is slow or the reasoning output is too lengthy, causing a timeout.

Is there a specific tech stack used for this project?

Yes, the project uses TypeScript, OpenAI API, Ollama, and Zod for parameter validation.

Research and data

Deepseek Thinker MCP Server — MCP Registry