PromptForge
Back to list
开发工具

MCP 协议 Python 服务端快速开发模板

使用FastMCP框架快速搭建Model Context Protocol服务端,包含工具定义、资源管理、认证配置和部署方案的完整开发指南

5 views4/22/2026

You are an expert MCP (Model Context Protocol) server developer specializing in FastMCP, the Pythonic framework for building MCP servers.

Task

Help me build a production-ready MCP server for the following use case:

  • Domain: [describe your data source or service, e.g., "internal knowledge base", "database queries", "API gateway"]
  • Target clients: [Claude Desktop / Cursor / custom agent]
  • Key operations needed: [list 3-5 operations]

Generate the Following

1. Project Structure

my-mcp-server/
├── server.py          # Main FastMCP server
├── tools/             # Tool implementations
├── resources/         # Resource providers
├── prompts/           # Prompt templates
├── auth.py            # Authentication
├── pyproject.toml     # Dependencies
└── README.md

2. Server Implementation

Using FastMCP decorators, implement:

  • @mcp.tool() — Define tools with proper input schemas, descriptions, and error handling
  • @mcp.resource() — Expose data resources with URI templates
  • @mcp.prompt() — Reusable prompt templates
  • Proper logging and error boundaries

3. Authentication & Security

  • API key validation middleware
  • Rate limiting per client
  • Input sanitization for all tool parameters
  • Audit logging

4. Testing Strategy

  • Unit tests for each tool
  • Integration tests with MCP Inspector
  • Load testing configuration

5. Deployment Options

  • Local stdio mode (for Claude Desktop)
  • HTTP/SSE mode (for remote clients)
  • Docker containerization
  • Cloud deployment (Railway / Fly.io / AWS Lambda)

6. Client Configuration

Generate the correct claude_desktop_config.json or client configuration snippet.

Constraints

  • Use Python 3.11+
  • Follow FastMCP best practices
  • Include comprehensive docstrings
  • Handle edge cases gracefully
  • Return structured responses

Please generate the complete, copy-paste ready implementation.