使用说明
MCP演示
这是一个基本的MCP服务器实现，它公开了数据和操作，供连接的大型语言模型使用。
ChatGPT的示例用法
启动服务器后，向ChatGPT提供以下说明
- You are connected to a remote tool MCP Demo.
- I will describe the usage of functions it contains, the schemas for each function's arguments, and the expected return format.
1. resources/list: Get a list of available resources. Takes no arguments, returns an array of resources with URIs and MIME types
2. tools/list: Get a list of available tools. Takes no arguments, returns an array of tool names
3. tools/call : Use a tool. Required parameters: 'name': The string name of the tool you want to use, 'params': A dictionary representing the tool's arguments
4. prompts/get: Retrieve a prompt. Required parameter: 'name': The string name of the prompt you want to retrieve, returns a string of the prompt text
Thank you, and welcome to MCP Demo
开始使用
通过Smithery安装
通过以下方式自动安装Claude Desktop的MCP演示 史密瑟里:
npx -y @smithery/cli install @PawNzZi/aidaily --client claude
资源
MCP演示包括可以查询的示例资源：
resources = [
{"name": "Hello World", "uri": "text://hello-world", "mimeType": "text/plain"},
{"name": "Introduction to Large Language Models", "uri": "text://introduction-to-llms", "mimeType": "text/plain"}
]
A line from A Introduction to Large Language Models
1. History: Large Language Models (LLMs) trace their roots to early research in artificial neural networks
返回的JSON编码响应 tools/list 调用应该看起来像：
{"jsonrpc":"2.0","id":1,"result":[{"name":"Example Tool","input":"Prompt","output":"Reply"}]}
目前只有一小部分行动和数据可用，但我们计划在未来用更令人兴奋的功能来扩展它！
安装
确保系统上安装了python，然后执行以下操作：
git clone THIS_REPOSITORY
pip install .
使用以下命令设置.env API_KEY="YOUR_KEY"
跑
使用以下命令运行服务器
python3 -m mcp_server
服务器监听端口8080
