Skip to main content
Glama
enesjakozh
moyu6027

DeepSeek MCP Server

by moyu6027
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Remote

🧠 DeepSeek MCP 服务器

DeepSeek MCP 服务器

🚀 功能

通过集成 DeepSeek R1 的高级推理引擎**，增强 Claude 的推理能力**。该服务器利用 deepseek r1 模型的推理能力，使 Claude 能够处理复杂的推理任务。

  • DeepSeek R1（大脑）充当高级推理规划器：

    • 规划多步骤逻辑分析策略

    • 构建认知框架

    • 评估信心和不确定性

    • 监控推理质量

    • 检测边缘情况和偏见

  • 克劳德（执行者）实施推理计划：

    • 执行结构化分析

    • 实施计划的战略

    • 提供最终答复

    • 处理用户交互

    • 管理系统集成

Related MCP server: Deepseek R1 MCP Server

🚀 功能

高级推理能力

  • 支持复杂的多步骤推理任务。

  • 旨在精确、高效地生成深思熟虑的回应。

  • 使用无问芯穹顶的API

完整的安装指南

先决条件

  • Python 3.12 或更高版本

  • uv包管理器

  • INFINI_API_KEY For DeepSeek（在无问芯穹注册）

  1. 克隆存储库

    git clone https://github.com/moyu6027/deepseek-MCP-server.git
cd deepseek-MCP-server

  2. 确保紫外线已设置

    • Windows ：在 PowerShell 中运行以下命令：

      powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

    • Mac ：运行以下命令：

      curl -LsSf https://astral.sh/uv/install.sh | sh

  3. 创建虚拟环境

    uv venv
source .venv/bin/activate

  4. 安装依赖项

    uv add "mcp[cli]" httpx

  5. 设置 API 密钥

    echo "INFINI_API_KEY=your_key_here" > .env

  6. 安装服务器

    mcp install server.py -f .env

  7. 配置 MCP 服务器编辑claude_desktop_config.json文件以包含以下配置：

    {
    "mcpServers": {
        "deepseek-mcp": {
            "command": "uv",
            "args": [
                "--directory",
                "PATH_TO_DEEPSEEK_MCP_SERVER",
                "run",
                "server.py"
            ]
        }
    }
}

  8. 运行服务器

    uv run server.py

🛠 使用方法

启动服务器

与 Claude Desktop 一起使用时，服务器会自动启动。请确保 Claude Desktop 已配置为检测 MCP 服务器。

示例工作流程

  1. 克劳德收到一个需要高级推理的查询。

  2. 该查询被转发到 DeepSeek R1 进行处理。

  3. DeepSeek R1 返回包裹在<ant_thinking>标签中的结构化推理。

  4. 克劳德将推理融入到最终的回应中。

📄 许可证

本项目遵循 MIT 许可证。详情请参阅LICENSE文件。

Available Tools

1 tool
reasonB
Process a query using DeepSeek's R1 reasoning engine and prepare it for integration with DeepSeek V3 or claude.

DeepSeek R1 leverages advanced reasoning capabilities that naturally evolved from large-scale 
reinforcement learning, enabling sophisticated reasoning behaviors. The output is enclosed 
within `<ant_thinking>` tags to align with V3 or Claude's thought processing framework.

Args:
    query (dict): Contains the following keys:
        - context (str): Optional background information for the query.
        - question (str): The specific question to be analyzed.

Returns:
    str: The reasoning output from DeepSeek, formatted with `<ant_thinking>` tags for seamless use with V3 or Claude.
ParametersJSON Schema
NameRequiredDescriptionDefault
queryYes

Output Schema

ParametersJSON Schema
NameRequiredDescription
resultYes

TDQS

B3.4/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations are provided, so the description carries the full burden. It mentions the tool 'leverages advanced reasoning capabilities' and outputs formatted text, but fails to disclose critical behavioral traits such as rate limits, error handling, authentication requirements, or performance characteristics. The description adds some context about the reasoning engine but leaves significant gaps for a tool with potential computational costs.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is appropriately sized and front-loaded, with the core purpose stated in the first sentence. Additional sentences provide useful context about the reasoning engine and output formatting. There is minor redundancy in mentioning 'V3 or Claude' twice, but overall, it's efficient and well-structured.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's moderate complexity (1 parameter with nested structure), no annotations, and an output schema that exists (though not detailed here), the description is reasonably complete. It explains the purpose, parameter semantics, and output format, though it could improve by addressing behavioral aspects like error cases or integration specifics.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The description adds substantial meaning beyond the input schema, which has 0% description coverage and only specifies a generic object. It details that the 'query' parameter is a dict with 'context' (optional background) and 'question' (specific question) keys, clarifying the expected structure and semantics. This compensates well for the schema's lack of documentation.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose: 'Process a query using DeepSeek's R1 reasoning engine and prepare it for integration with DeepSeek V3 or claude.' It specifies the verb ('process'), resource ('query'), and technology ('DeepSeek's R1 reasoning engine'), but since there are no sibling tools, it cannot demonstrate differentiation from alternatives.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies usage by mentioning integration with 'V3 or Claude's thought processing framework,' suggesting it's for preparing reasoning outputs for those systems. However, it lacks explicit guidance on when to use this tool versus alternatives (e.g., direct API calls or other reasoning engines) and does not specify prerequisites or exclusions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 1 tool updatev1.0.0
    • First observedreason
Deploy Server
Shareable, User-Owned Context BackendPeople, applications, and agents read, write, and coordinate together in one context lake. Reach it over MCP, a CLI, a Python SDK, or REST.fulcradynamics.comAd

TDQS

B3.4/5.0

Scored across 1 tool

Disambiguation5/5

With only one tool named 'reason', there is no possibility of ambiguity or overlap with other tools. The tool has a single, clearly defined purpose of processing queries through DeepSeek's reasoning engine.

Naming Consistency5/5

A single tool cannot demonstrate inconsistency in naming patterns. The tool name 'reason' follows a clear verb-based convention that directly describes its function.

Tool Count2/5

One tool is too few for a server that appears to interface with a complex reasoning engine. While the tool is well-described, a single tool surface severely limits the server's capabilities and suggests an incomplete implementation for the domain.

Completeness2/5

The server has a significant gap in functionality. A reasoning engine server should offer more than just query processing - there are no tools for configuration, status checking, result formatting options, or other typical operations expected from such a service.

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

Appeared in Searches