MCP Server GuideMCP Server 指南

What Is an MCP Server?什么是 MCP 服务器？

An MCP server is a capability provider in the Model Context Protocol that exposes capabilities to an MCP client. A server can publish tools, resources, and prompts through JSON-RPC, while the transport layer carries those messages over stdio or Streamable HTTP. The host—not the server—decides how capabilities enter the model context and when user approval is required.MCP 服务器是模型上下文协议中的功能提供者，向 MCP 客户端公开功能。服务器可以通过 JSON-RPC 发布工具、资源和提示，而传输层则通过 stdio 或 Streamable HTTP 承载这些消息。主机（而不是服务器）决定功能如何进入模型上下文以及何时需要用户批准。

An MCP server is the capability provider in the Model Context Protocol. While MCP clients (inside AI applications like Claude Desktop or Cursor) initiate connections, servers respond to those connections by exposing the tools, resources, and prompts available to connected clients.MCP 服务器是模型上下文协议中的能力提供者。当 MCP 客户端（在 Claude Desktop 或 Cursor 等 AI 应用程序内部）发起连接时，服务器通过向连接的客户端公开可用的工具、资源和提示来响应这些连接。

MCP has a required base protocol and lifecycle, while individual features are negotiated during initialization. A server may supportMCP 具有必需的基本协议和生命周期，而各个功能则在初始化期间协商。服务器可能支持 tools/list工具/列表 and和 tools/call工具/调用, resources/list and resources/read资源/列表和资源/阅读, or， 或者 prompts/list and prompts/get提示/列表和提示/获取. Do not assume every server exposes all three primitives.。不要假设每个服务器都公开所有三个原语。

The ecosystem includes official SDKs, reference implementations, vendor-hosted remote servers, local community packages, and public registries. A listing or repository star count is discovery evidence—not a security certification. Start with the official该生态系统包括官方 SDK、参考实现、供应商托管的远程服务器、本地社区包和公共注册表。列表或存储库星级计数是发现证据，而不是安全认证。从官方开始 MCP server conceptsMCP 服务器概念 and和 architecture overview架构概述, then verify the exact server artifact you plan to execute.，然后验证您计划执行的确切服务器工件。

Production safety is shared across participants. Servers must validate inputs, enforce access control, rate-limit calls, and sanitize outputs; clients should show sensitive inputs, request confirmation, apply timeouts, validate results, and retain audit evidence. See the official安全生产由参与者共享。服务器必须验证输入、实施访问控制、限制调用速率并清理输出；客户应显示敏感输入、请求确认、应用超时、验证结果并保留审计证据。看官方 MCP tools security requirementsMCP 工具安全要求 and和 authorization specification授权规范.

MCP Server vs MCP Client: The ArchitectureMCP 服务器与 MCP 客户端：架构

The MCP architecture separates concerns clearly: servers provide capabilities, clients consume them, and hosts run one or more clients. This stateless design is intentional — servers don't track which clients connect to them, enabling composability across any MCP-compatible application.MCP 架构清楚地划分了关注点：服务器提供功能，客户端使用它们，主机运行一个或多个客户端。这种无状态设计是有意为之的——服务器不会跟踪哪些客户端连接到它们，从而实现了跨任何 MCP 兼容应用程序的可组合性。

Key architectural property:主要建筑属性： Clients don't know which other clients are connected to the same server, and servers don't track which clients are using them. A client sends a tools/list request and receives a list of available tools. The server simply responds to requests — it doesn't maintain session state or client identity. This means Claude Desktop, Cursor, Cline, and any other MCP-compatible client can all connect to the same MCP server without the server needing special configuration.客户端不知道哪些其他客户端连接到同一服务器，服务器也不跟踪哪些客户端正在使用它们。客户端发送工具/列表请求并接收可用工具的列表。服务器只是响应请求——它不维护会话状态或客户端身份。这意味着 Claude Desktop、Cursor、Cline 和任何其他 MCP 兼容客户端都可以连接到同一 MCP 服务器，而无需服务器进行特殊配置。

For developers building AI systems, the practical implication: your choice of MCP server is independent of your choice of MCP client. All MCP-compatible applications can connect to the same servers, because the MCP specification standardizes the communication interface.对于构建 AI 系统的开发人员来说，实际意义是：您对 MCP 服务器的选择与 MCP 客户端的选择无关。所有 MCP 兼容的应用程序都可以连接到相同的服务器，因为 MCP 规范标准化了通信接口。

MCP Server Selection MatrixMCP 服务器选择矩阵

Use this matrix to choose a server pattern by trust boundary and workload. Confirm every field against the exact release you deploy; names, schemas, transports, and permissions can change independently.使用此矩阵按信任边界和工作负载选择服务器模式。根据您部署的确切版本确认每个字段；名称、模式、传输和权限可以独立更改。

Validation rule:验证规则： Measure initialization, discovery, authorization, execution, retries, and result validation separately. Never reuse a latency number from another region, provider, payload, account tier, or concurrency level.分别测量初始化、发现、授权、执行、重试和结果验证。切勿重复使用来自其他区域、提供商、负载、帐户层或并发级别的延迟数字。

Eight MCP Server Examples and What to Verify八个 MCP 服务器示例以及要验证的内容

These examples illustrate different capability and trust patterns. Treat every count, schema, transport, and permission as release-specific. Labels marked这些示例说明了不同的能力和信任模式。将每个计数、模式、传输和权限视为特定于版本的。标签标记 Verified已验证 indicate where the publisher offers an official implementation or documentation—not a blanket security approval.指明发布者在何处提供官方实现或文档，而不是一揽子安全批准。

1. GitHub MCP1.GitHubMCP Verified已验证

2. Figma MCP2. Figma MCP

3. Linear MCP3. 线性MCP

4. Sentry MCP4.哨兵MCP Verified已验证

5. Context75. 背景7

6. Brave Search MCP6.勇敢搜索MCP Verified已验证

7. Supabase MCP7.Supabase MCP

8. Notion MCP8.MCP概念

MCP Server Examples: How They Work in PracticeMCP 服务器示例：它们在实践中如何工作

Understanding how an MCP server works under the hood helps you evaluate servers, debug connection issues, and build custom servers. Below is a minimal FastMCP server implementation and the JSON-RPC flow that powers tool discovery.了解 MCP 服务器的工作原理有助于您评估服务器、调试连接问题以及构建自定义服务器。下面是一个最小的 FastMCP 服务器实现和支持工具发现的 JSON-RPC 流程。

Building a Minimal MCP Server with the Python SDK使用 Python SDK 构建最小的 MCP 服务器

# Minimal MCP server with FastMCP
# Source: official MCP Python SDK documentation
# Pin and test the current SDK release before deployment

from mcp.server.fastmcp import FastMCP

mcp = FastMCP("my-server")

@mcp.tool()
def search_docs(query: str) -> str:
    """Search documentation for a query string."""
    return f"Results for: {query}"

@mcp.tool()
def get_doc_content(doc_id: str) -> str:
    """Retrieve the full content of a document by ID."""
    # Implementation would fetch from your doc store
    return f"Content of {doc_id}"

# Run: python my_server.py
# Or: fastmcp dev my_server.py (hot reload for development)
# Connect: Configure in claude_desktop_config.json as stdio

The FastMCP decorator-based approach keeps server code minimal. Each decorated function automatically becomes a tool with its name from the function name, description from the docstring, and input schema from type annotations.基于 FastMCP 装饰器的方法使服务器代码保持最少。每个修饰函数都会自动成为一个工具，其名称来自函数名称，描述来自文档字符串，输入模式来自类型注释。

JSON-RPC Request/Response FlowJSON-RPC 请求/响应流程

Tool discovery and invocation use JSON-RPC 2.0 over stdio or Streamable HTTP:工具发现和调用通过 stdio 或 Streamable HTTP 使用 JSON-RPC 2.0：

// Step 1: Client discovers available tools
// Client → Server (via stdio stdin or HTTP POST)
{"jsonrpc": "2.0", "id": 1, "method": "tools/list", "params": {}}

// Step 2: Server returns tool definitions
// Server → Client (via stdio stdout or HTTP response)
{"jsonrpc": "2.0", "id": 1, "result": {
  "tools": [
    {"name": "search_docs", "description": "Search documentation for queries", "inputSchema": {"type": "object", "properties": {"query": {"type": "string"}}}}
  ]
}}

// Step 3: Client invokes a tool
{"jsonrpc": "2.0", "id": 2, "method": "tools/call", "params": {
  "name": "search_docs", "arguments": {"query": "authentication best practices"}
}}

// Step 4: Server returns results
{"jsonrpc": "2.0", "id": 2, "result": {
  "content": [{"type": "text", "text": "Results for: authentication best practices"}]
}}

Every request includes a每个请求都包含一个 jsonrpc version, an版本，一个 id for correlation, a对于相关性，a method name, and姓名，以及 params. Responses include the same。回复包括相同内容 id and a和一个 result or或者 error. This simplicity makes MCP servers easy to implement and debug — no custom protocol parsing or state management required.。这种简单性使得 MCP 服务器易于实现和调试——无需自定义协议解析或状态管理。

Production MCP Server Workflow and Trust Boundaries生产 MCP 服务器工作流程和信任边界

A production flow should discover candidates, inspect provenance and schemas, authorize the narrowest action, call with validated arguments, and validate the result against its source. Keep the host, MCP server, authorization server, and downstream service boundaries explicit in logs and incident response.生产流程应该发现候选者，检查出处和模式，授权最窄的操作，使用经过验证的参数进行调用，并根据其来源验证结果。在日志和事件响应中明确主机、MCP 服务器、授权服务器和下游服务边界。

Decision framework:决策框架：

QVeris can serve as an optional capability discovery and routing layer in this workflow (QVeris 可以充当此工作流程中的可选功能发现和路由层（QVeris CLI documentationQVeris CLI 文档). MCP still defines the protocol, the host controls context and approvals, each MCP server enforces authorization, and downstream providers remain responsible for their own data and operations.）。 MCP 仍然定义协议，主机控制上下文和批准，每个 MCP 服务器强制授权，下游提供商仍然对自己的数据和操作负责。

Deployment Options for MCP ServersMCP 服务器的部署选项

MCP servers can be deployed in three patterns, each with different trade-offs for security, scalability, and operational overhead.MCP 服务器可以以三种模式部署，每种模式在安全性、可扩展性和运营开销方面都有不同的权衡。

Local Process (stdio)本地进程（stdio）

Simplest deployment: run the MCP server as a local process and connect via stdio. Works well for development, CLI tools, and single-user workflows. Claude Code and other CLI-based clients typically use stdio transport. Configuration is minimal — point your MCP client at the process command. The downside: the server is tied to the local machine, preventing team sharing and remote access.最简单的部署：将 MCP 服务器作为本地进程运行并通过 stdio 连接。非常适合开发、CLI 工具和单用户工作流程。 Claude Code 和其他基于 CLI 的客户端通常使用 stdio 传输。配置最少 - 将 MCP 客户端指向进程命令。缺点：服务器与本地计算机绑定，阻止团队共享和远程访问。

Remote Service (Streamable HTTP)远程服务（流式 HTTP）

For production deployments where multiple clients need to connect, deploy the MCP server as a remote HTTP service. Requires authentication (API keys, OAuth, or mTLS) and connection health monitoring. Per对于需要连接多个客户端的生产部署，请将 MCP 服务器部署为远程 HTTP 服务。需要身份验证（API 密钥、OAuth 或 mTLS）和连接运行状况监控。每 Toolradar's April 2026 endpoint scan of 2,181 serversToolradar 于 2026 年 4 月对 2,181 台服务器进行的端点扫描, 52% of remote MCP servers become unresponsive within 30 days due to missing health checks and auto-restart configuration. Production deployments must implement: health check endpoints, automatic restart on failure, and connection timeout handling.，52% 的远程 MCP 服务器由于缺少运行状况检查和自动重启配置而在 30 天内变得无响应。生产部署必须实现：健康检查端点、失败时自动重启以及连接超时处理。

Managed Platform托管平台

Managed MCP platforms like托管 MCP 平台，例如 Apigene芹菜基因, Prefect长官, and AWS Bedrock handle authentication, monitoring, scaling, and security.和 AWS Bedrock 处理身份验证、监控、扩展和安全性。 CData's MCP infrastructure market analysis (Q1 2026)CData 的 MCP 基础设施市场分析（2026 年第一季度） projects the managed hosting segment at $10.4B with 24.7% CAGR.预计管理托管细分市场为 $10.4B，复合年增长率为 24.7%。

// Example: Claude Desktop config connecting a meta-server
// macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
// Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "qveris": {
      "command": "npx",
      "args": ["-y", "@qverisai/mcp"],
      "env": {
        "QVERIS_API_KEY": "your-api-key"
      }
    }
  }
}

// Alternative: connecting a single-purpose server directly
// See each server's repository for its specific config format

After configuring, restart your MCP client. The server's tools become available through the standard MCP tool discovery flow.配置完成后，重新启动 MCP 客户端。服务器的工具通过标准 MCP 工具发现流程变得可用。

FAQ: MCP Servers常见问题解答：MCP 服务器

Related Guides相关指南