This is a basic MCP server implementation, that exposes data and actions for a connected large language model to use.
Give the following instructions to ChatGPT after starting the server
- 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
To install MCP Demo for Claude Desktop automatically via Smithery:
npx -y @smithery/cli install @PawNzZi/aidaily --client claudeThe MCP Demo includes example resources that can be queried:
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 an Introduction to Large Language Models
1. History: Large Language Models (LLMs) trace their roots to early research in artificial neural networks
The returned JSON-encoded response of the tools/list call should look something like:
{"jsonrpc":"2.0","id":1,"result":[{"name":"Example Tool","input":"Prompt","output":"Reply"}]}
Currently only a small set of actions and data is available but we plan to expand this with more exciting capabilities in the future!
Ensure python is installed on the system and then do the following:
git clone THIS_REPOSITORY
pip install .
Setup the .env with an API_KEY="YOUR_KEY"
Run the server with
python3 -m mcp_server
The server listens on port 8080