mcp-hands-on-with-agentic-ai

概述

这个仓库提供了一种实践方法来学习和实现模型上下文协议(MCP)。它包含了多个Python和TypeScript的MCP服务器示例，展示了如何扩展AI代理的功能，包括文本分析、天气数据检索、项目文档生成和GitHub模型比较等。此外，还包含用于创建新MCP服务器的模板，对于希望将MCP集成到应用程序或构建新的智能AI工具的开发者来说非常有价值。

试试问 AI

装完之后，这里有 5 个你可以让 AI 做的事：

• count_characters

  Count the number of characters in a given text

  试试:Count characters in this text试试:How many characters are in 'Hello World?'
• count_words

  Count the number of words in a given text

  试试:Count words in this paragraph试试:How many words are in this document?
• get_current_weather

  Get current weather information for a location

  试试:What's the weather like in London today?试试:Show me current weather conditions
• get_forecasted_weather

  Get forecasted weather information for a location

  试试:What's the weather forecast for Paris this week?试试:Show me weather forecast for next 7 days
• summarize_project

  Summarize a project

  试试:Summarize this project试试:Give me a brief overview of this codebase
• generate_readme

  Generate a comprehensive README.md document

  试试:Generate a README for this project试试:Create documentation for this software
• list_github_models

  List all available GitHub Models

  试试:List all GitHub models试试:Show me available GitHub models
• compare_github_models

  Compare GitHub models

  试试:Compare gpt-4 and claude-3试试:Show differences between these GitHub models
• run_model_comparison

  Run completion comparisons between models

  试试:Compare completions from gpt-4 and claude-3试试:Show me how these models respond to this prompt

说明：Tool names were inferred from folder descriptions and README sections. No explicit tool documentation was provided in the main README.

可对比工具

安装

FAQ

什么是模型上下文协议(MCP)？

MCP是一个通用协议，允许开发人员通过一致地提供上下文、工具和指令来为LLM添加代理行为。

这个仓库支持哪些编程语言？

这个仓库提供了Python和TypeScript MCP服务器的示例和模板。

mcp-hands-on-with-agentic-ai 对比