MCP Server 8im

An MCP server project for experimenting with LLM tools, providing a variety of local and network tool integrations.

Developer tools Artificial intelligence chatbots #LLM tools #MCP protocol #API integration #Experimental project .Python

rating : 2.5 points

downloads : 6.9K

update time : 2025-04-23

Open Site

What is the MCP Server?

The MCP server is a tool integration platform specifically designed for large language models (LLMs). It enables LLMs to call various utility tools and API services through the Model Context Protocol.

How to use the MCP Server?

You can start the server via a simple command line and then interact with the integrated tools through a chat application or direct API calls.

Applicable scenarios

Suitable for development scenarios that require extending LLM capabilities, such as data analysis, information retrieval, and automated tasks.

Main features

Basic tools

Provides basic calculation functions such as addition, calculator, and BMI calculation

Search tools

Integrates search services such as DuckDuckGo, Wikipedia, and ArXiv

API integrations

Supports various API services such as OpenWeatherMap, NewsAPI, and GitHub

File processing

Provides long - task processing and image thumbnail generation functions

Advantages

Rich tool integration to meet various needs in one stop

Supports both local and remote running modes

Provides an intuitive web interface to view and manage tools

Limitations

Some API tools require additional configuration of keys

Currently does not support the default integration of all LLM models

Primarily for experimental purposes, not recommended for production environments

How to use

Install dependencies

First, you need to install the uv tool and synchronize dependencies

Start the server

Choose to start in local development mode or network server mode

Configure environment variables

Set environment variables for tools that require API keys

Use tools

Use the integrated tools through a chat application or direct API calls

Usage examples

Weather query

Get weather information for a specified city through the OpenWeatherMap API

Academic search

Search for papers related to artificial intelligence on arXiv

GitHub operations

Get the content of files in a GitHub repository

Frequently Asked Questions

Why can't some tools be used?

How to view the list of available tools?

Does it support custom tools?

Related resources

Ollama official website

A tool for running large language models locally

OpenWeatherMap API documentation

Documentation for the weather API service

GitHub API documentation

GitHub REST API reference

🚀 MCP Server

An MCP server for experimenting with LLM tools

This project aims to understand the MCP server, its protocols, and its usage in LLMs. Reuse is not recommended!

🚀 Quick Start

📦 Installation

Dependencies

Install 'uv'
Run uv sync

Unit Tests

uv run pytest

🚀 Starting the Server

Local/Standard Input and Output

uv run mcp dev server.py

(.venv) ➜  mcp-server git:(main) ✗ uv run mcp dev server.py
Starting MCP checker...
Proxy server listening on port 3000

🔍 MCP checker is up and running at http://localhost:5173 🚀

As a Web Server

This allows remote clients to connect to the MCP server.

uv run fastmcp run server.py --transport sse

The server will be available at http://localhost:8000.

💻 Usage Examples

Viewing Tools

Available Tools

Tool	Description	Backend Service	Required Configuration
add	Simple addition tool	Local computation	None
calculator	Calculate mathematical expressions	Local computation	None
calculate_bmi	Calculate body mass index	Local computation	None
echo	Return input text unchanged	Local computation	None
long_task	Process files and track progress	Local file system	None
duckduckgo_search	Perform web search using DuckDuckGo	DuckDuckGo HTML endpoint	None
wikipedia_search	Search Wikipedia articles	Wikipedia API	None
fetch_weather	Get current weather based on location	OpenWeatherMap API	`OPENWEATHER_API_KEY`
openmeteo_forecast	Get detailed weather forecast	Open-Meteo API	None
news_search	Search recent news articles	NewsAPI	`NEWSAPI_KEY`
tavily_search	AI-driven web search	Tavily API	`TAVILY_API_KEY`
arxiv_search	Search academic papers	arXiv API	None
github_get_file	Get file content from a GitHub repository	GitHub API	`GITHUB_TOKEN`
github_list_issues	List issues in a repository	GitHub API	`GITHUB_TOKEN`
github_create_issue	Create a new issue in a repository	GitHub API	`GITHUB_TOKEN`
github_list_pull_requests	List pull requests in a repository	GitHub API	`GITHUB_TOKEN`
github_caveats	Note that it currently does not work with the default model... Work in progress!

Environment Variable Configuration

OPENWEATHER_API_KEY - OpenWeatherMap API key required to get weather data
NEWSAPI_KEY - NewsAPI API key required to get news data
TAVILY_API_KEY - Tavily API key required to get web search results
GITHUB_TOKEN - Token required to access the GitHub API

Example Chat Application

To use the example chat application:

Clone this repository.
Install dependencies: pip install -r requirements.txt
Run the chat application: python chat_app.py

Features

The chat application automatically uses MCP tools when needed.
All interactions are logged in the specified LOG_FILE.
Tools will be invoked when the LLM deems them helpful for answering questions.
Tool parameters will be automatically populated based on the LLM's understanding of the query.