Deepseek R1 MCP Server
@66julienmartin
A Model Context Protocol (MCP) server implementation connecting Claude Desktop with DeepSeek's language models (R1/V3)
Overview
What is Deepseek R1 MCP Server?
An MCP server implementation that integrates the Deepseek R1 language model with Claude Desktop. Deepseek R1 is optimized for reasoning tasks and has an 8192‑token context window. This server is aimed at developers who want to use Deepseek’s models through the MCP protocol.
How to use Deepseek R1 MCP Server?
Clone the repository, run npm install, copy .env.example to .env and add your Deepseek API key, then build with npm run build. In Claude Desktop’s configuration, add an MCP server entry with command node and arguments pointing to the built index.js file, plus the DEEPSEEK_API_KEY environment variable.
Key features of Deepseek R1 MCP Server
- Advanced text generation with 8192‑token context window
- Configurable parameters (max_tokens, temperature)
- Robust error handling with detailed messages
- Full MCP protocol support
- Claude Desktop integration
- Support for both DeepSeek‑R1 and DeepSeek‑V3 models
Use cases of Deepseek R1 MCP Server
- Code generation and mathematical problem solving
- Data cleaning and data analysis tasks
- General conversational chat and dialogue
- Language translation
- Creative writing and poetry generation
FAQ from Deepseek R1 MCP Server
How do I switch from DeepSeek‑R1 to DeepSeek‑V3?
Modify the model field in src/index.ts. Use "deepseek-reasoner" for R1 (default) or "deepseek-chat" for V3.
What are the runtime requirements?
Node.js v18 or higher, npm, Claude Desktop, and a valid Deepseek API key.
What is the default temperature and how should I adjust it?
The default temperature is 0.2. For coding and math use 0.0, for data analysis 1.0, for general conversation and translation 1.3, for creative writing 1.5.
Which parameters can I pass in an API call?
You can set prompt, max_tokens (up to 8192), and temperature (0.0 – 2.0). See the API usage section in the README.
What kind of error messages does the server provide?
Detailed errors for API authentication failures, invalid parameters, rate limiting, and network issues.
