广告合作
MCP.so
登录

DeepSeek MCP Server

@DMontgomery40

代码仓库主页

基本信息

分类

其他

许可证

MIT

传输方式

stdio

作者

DMontgomery40

关于 DeepSeek MCP Server

Model Context Protocol server for DeepSeek's advanced language models

#mcp#r1#deepseek-chat#deepseek-api#model-context-protocol#deepseek-v3#deepseek-r1

DeepSeek MCP Server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "deepseek-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "deepseek-mcp-server"
      ],
      "env": {
        "DEEPSEEK_API_KEY": "REPLACE_WITH_DEEPSEEK_KEY"
      }
    }
  }
}

DeepSeek MCP Server 工具

7

Primary DeepSeek V4 chat tool for single-turn and multi-turn generation. Defaults to `deepseek-v4-flash`; use `deepseek-v4-pro` for higher-capability reasoning. Provide either `message` (simple single user turn) or `messages` (full chat history); if both are provided, `messages` is used. Thinking mode is enabled by DeepSeek by default; pass `thinking:{type:"disabled"}` for non-thinking mode, and use `reasoning_effort:"low"|"high"|"max"` when thinking is enabled. Use `conversation_id` to persist context across calls and `clear_conversation=true` to reset stored state before sending the next turn. Set `include_raw_response=true` only for debugging because it returns the full provider payload.

DeepSeek V4 Pro FIM completion tool for prompt/suffix fill-in-the-middle workflows. Defaults to `deepseek-v4-pro`. Use this when you need raw completion text instead of chat message formatting. Set `include_raw_response=true` only when you need the full provider payload for debugging.

Create a stateless DeepSeek V4 response using the native OpenAI-compatible Responses API. Defaults to `deepseek-v4-flash`. Provide `input`, `instructions`, or both. Use `reasoning.effort` for thinking control, `tools` for function or server-side web-search tools, and `stream=true` for semantic SSE aggregation. This tool does not persist provider-side response state; send the full input history for multi-turn work. Set `include_raw_response=true` only for debugging because it returns the full provider payload.

List available DeepSeek models for model selection and validation. This tool takes no parameters. Use it before passing an explicit model ID to generation tools.

Return the current DeepSeek account balance and availability status. This tool takes no parameters and is read-only. Use it for account health checks when diagnosing provider-side failures.

Delete stored in-memory chat history for a `conversation_id`. Use this when you want to keep the same ID but start a fresh thread. This only affects server-side memory in the current MCP process.

List all conversation IDs currently stored in this MCP process memory. This tool takes no parameters and does not call the DeepSeek API. Useful for debugging conversation persistence behavior.

概览

What is DeepSeek MCP Server?

A Model Context Protocol server that exposes DeepSeek’s V4 API (chat completions, FIM completions, model listing, account balance) as MCP tools. Designed for developers using MCP-compatible hosts like Claude Code, Cursor, or Codex CLI.

How to use DeepSeek MCP Server?

Run locally via npx deepseek-mcp-server or Docker, setting the DEEPSEEK_API_KEY environment variable. Alternatively, connect to the hosted remote at https://deepseek-mcp.ragweld.com/mcp using a bearer token. Transport is stdio by default; HTTP can be configured via environment variables.

Key features of DeepSeek MCP Server

  • Chat completions with deepseek-v4-flash and deepseek-v4-pro
  • FIM (fill-in‑the‑middle) completion with deepseek-v4-pro
  • List models and check account balance
  • In‑memory conversation history with reset and list tools
  • Supports streaming, thinking, reasoning effort, JSON output, and function tools
  • Official MCP Registry listing and OCI container image

Use cases of DeepSeek MCP Server

  • Integrate DeepSeek V4 chat into a Claude Code or Cursor workflow
  • Automate FIM completions for code generation within MCP agents
  • Monitor account usage and available models programmatically
  • Build custom AI tools that rely on DeepSeek’s latest API

常见问题

1

What tools does DeepSeek MCP Server expose?

It provides `chat_completion`, `completion`, `list_models`, `get_user_balance`, `reset_conversation`, and `list_conversations`. Only documented DeepSeek V4 API endpoints are supported.

2

How do I authenticate when using the hosted remote?

Set the environment variable `DEEPSEEK_MCP_AUTH_TOKEN` and pass it as a bearer token in the `Authorization` header. For local usage, set `DEEPSEEK_API_KEY`.

3

What runtime and environment variables are required?

Requires Node.js (to run via npx) and the `DEEPSEEK_API_KEY` environment variable. Optional variables include `DEEPSEEK_BASE_URL`, `DEEPSEEK_REQUEST_TIMEOUT_MS`, `MCP_TRANSPORT`, and `CONVERSATION_MAX_MESSAGES`.

4

Where is conversation data stored?

Conversation history is held in memory on the server. Use `reset_conversation` or `list_conversations` to manage it.

5

Does the server support image, video, or upload tools?

No. The server explicitly states it does not ship speculative image, video, or upload tools, nor automatic model substitution. Only documented chat, FIM, model listing, and balance endpoints are exposed.

评论

其他 分类下的更多 MCP 服务器

Sequential Thinking

modelcontextprotocol

Model Context Protocol Servers

88.2K

Soar Flight Booking MCP

Soar Flight Booking MCP is a remote MCP server available at https://mcp.soar.flights.

FastMCP v2 🚀

jlowin

🚀 The fast, Pythonic way to build MCP servers and clients.

26.0K

MaxKB

1Panel-dev

🔥 MaxKB is an open-source platform for building enterprise-grade agents. 强大易用的开源企业级智能体平台。

22.0K

Lemon.io

lemon-io

Request dedicated senior developers to work on your project, write job descriptions, and prep technical interviews — all without leaving your Claude chat. Just describe what you need, and Lemon.io MCP does the rest. 1–3

广告合作