Custom MCP Server

A customized MCP server project that integrates multiple tool functions such as Brave search, secure file system, and GitHub integration, providing rich operational capabilities for AI assistants.

Developer tools Artificial intelligence chatbots #AI tool integration #Multi - protocol support #Search function #File management .JavaScript

rating : 2 points

downloads : 12

update time : 2025-04-23

What is the Custom MCP Server?

The Custom MCP Server is an enhanced tool platform designed for AI assistants, providing extended functions for AI systems through the standardized MCP protocol. It's like installing a 'Swiss Army Knife' for AI, enabling it to perform complex tasks such as web search, file operations, and code management.

How to use the Custom MCP Server?

Simply install and connect your AI system (e.g., Claude), and you can immediately unlock all functions. The server provides user - friendly API interfaces, allowing AI assistants to use various tools as if they were built - in functions.

Applicable scenarios

Particularly suitable for AI assistant scenarios that need to handle complex tasks, such as: answering technical questions (can search for the latest information), assisting with code projects (can access GitHub), data analysis (can read and write files), etc.

Main features

Intelligent searchProvides web and local business information search capabilities through the Brave search engine to obtain the latest and most accurate network information.

Secure file managementSupports complete file system operations, including reading and writing files, creating directories, etc. All operations have strict security controls.

GitHub integrationInteracts directly with GitHub to perform development - related operations such as managing code repositories, viewing file contents, and pushing changes.

Step - by - step thinkingProvides a structured problem - solving framework to help AI systems handle complex problems more effectively.

File downloadDownloads single or multiple files from the Internet to the local area, expanding the data acquisition capabilities of AI.

Advantages and limitations

Advantages

Rich in functions: Integrates multiple practical tools to solve the extension needs of AI assistants in one stop.

Easy to integrate: Uses the standard MCP protocol and seamlessly connects with mainstream AI systems.

Safe and reliable: Sensitive functions such as file operations have strict security controls.

Open - source and flexible: Under the MIT license, it can be freely modified and extended as needed.

Limitations

Requires basic technical knowledge for initial setup and configuration.

Some functions (e.g., GitHub operations) require additional configuration of API keys.

Currently only supports running in the Node.js environment.

How to use

Get the server code

Clone the server code to your local environment.

Install dependencies

Install all software packages and dependencies required for running.

Configure the environment

Set the necessary API keys and configuration parameters (e.g., keys for Brave search and GitHub).

Start the server

Start both the main server and the MCP protocol server simultaneously.

Connect the AI system

Usage examples

Technical research assistantAI can help users conduct technical research and automatically search for the latest materials and information.

Code project assistantAI can directly interact with GitHub to help manage code projects.

Data analysis helperAI can read and analyze local data files to help users process data.

Frequently Asked Questions

What kind of technical background do I need to use this server?

Which AI systems does the server support?

Is file operation safe?

How to add custom functions?

What is the server performance like?

Related resources

GitHub repository

Project source code and latest updates.

MCP protocol documentation

Official documentation of the MCP protocol standard.

Brave search API

Guide to using the Brave search API.

Installation video tutorial

Step - by - step video guide for installation and configuration.

🚀 Custom MCP Server

A custom Machine Conversation Protocol (MCP) server integrating various tools for AI assistants, which can solve the problem of limited functionality of traditional servers and provide a more comprehensive and efficient service for AI assistants.

🚀 Quick Start

# Clone the repository
git clone https://github.com/degenhero/custom-mcp-server.git
cd custom-mcp-server

# Install dependencies and set up the environment
npm run install:all

# Start both servers
npm run start:both

The servers will run at:

MCP Server: http://localhost:3000
MCP Protocol Server: http://localhost:3001

💻 Usage Examples

Using with Claude

To use this custom MCP server with Claude, install it as an MCP server:

npx @anthropic-ai/install-mcp custom-mcp-server http://localhost:3001

After installation, Claude will be able to access all the features provided by the server.

🔧 Technical Details

Architecture

The project consists of two main components:

Custom MCP Server (index.js) - The core server that implements all tool functions.
MCP Protocol Server (mcp-protocol.js) - Adapts the custom server to the MCP protocol standard.

Available Features

Brave Search

brave_web_search - Perform web searches.
brave_local_search - Search for local businesses and locations.

File System Operations

read_file - Read file contents.
write_file - Create or update files.
edit_file - Perform line-based editing.
create_directory - Create directories.
list_directory - List directory contents.
etc.

GitHub Integration

create_repository - Create new repositories.
search_repositories - Search for repositories.
get_file_contents - Read files in repositories.
push_files - Push files to repositories.
etc.

Sequential Thinking

sequentialthinking - Solve problems step by step.

Download Tools

download_file - Download files from URLs.
download_multiple_files - Download multiple files simultaneously.

Configuration

Configuration is managed through environment variables in the .env file:

PORT=3000                      # Main server port
MCP_PROTOCOL_PORT=3001         # MCP protocol server port
BRAVE_API_KEY=your_key_here    # API key for Brave Search
GITHUB_TOKEN=your_token_here   # GitHub access token

Development

Adding New Features

To add new features:

Create handlers in the appropriate files in the functions directory.
Add function definitions to mcp-schema.json.
Update the routes in index.js.

Testing

You can test the MCP server directly:

curl -X POST http://localhost:3000/mcp \
  -H "Content-Type: application/json" \
  -d '{"functionName": "brave_web_search", "parameters": {"query": "test query"}}'