π NexusHub
About
NexusHub is a powerful Model Context Protocol (MCP) server that functions as a central connection point for AI workflows and tool integration.
Details
- Author
- webdevtodayjason
- GitHub stars
- 1
- Downloads
- 166
- Categories
- Other
Jump to
- Dual interface supporting both HTTP and stdio MCP protocols
- Modern dashboard with Protocol design system and dark mode
- Integrated services: Memory, GitHub, Brave Search
- Vector database for document ingestion and semantic search
- Secure API key and service configuration management
- Docker Compose deployment for easy setup
- Pre-configured advanced prompts for AI interactions
- Real-time health monitoring for all connected services
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
π NexusHubCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
Clone the repository, copy .env.example to .env and fill in API keys, then run docker-compose up -d. Configure Claude Desktop by adding the server to claude_desktop_config.json in either HTTP mode (using url) or STDIO mode (using command pointing to the wrapper script). Access the dashboard at http://localhost:8001/dashboard.
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"\ud83d\ude80 nexushub": {
"nexushub": {
"type": "http",
"url": "http://localhost:8001/mcp"
}
}
}
}
McpServers
{
"nexushub": {
"type": "http",
"url": "http://localhost:8001/mcp"
}
}
π NexusHub
Unified MCP Server for Claude AI Tools
<p align="center">

</p>
<p>
<strong>A powerful bridge between Claude AI and external services via Model Context Protocol.</strong><br>
Enhance Claude Code with filesystem access, database operations, vector search, GitHub integration, and more.
</p>
Demo β’
Documentation β’
Report Bug β’
Request Feature
</div>
---
β¨ Features
- π‘ Dual Interface - Supports both HTTP and stdio MCP protocols
- π
Modern Dashboard - Beautiful Protocol design system admin interface with dark mode by default
- π Integrated Services - Combines multiple MCP servers (Memory, GitHub, Brave Search)
- π Vector Database - Document ingestion and semantic search capabilities
- π API Management - Securely manage API keys and service configurations
- π³ Docker Integration - Easy deployment with Docker and Docker Compose
- π Advanced Prompts - Pre-configured prompts for effective AI interactions
- π Real-time Status - Monitor the health of all connected services
<div align="center">
<br>

<br>
<em>The elegant NexusHub dashboard interface with Protocol design system</em>
<br>
</div>
ποΈ Architecture
NexusHub uses a Node.js backend with Express for the HTTP interface and native stdio handling for the CLI interface. The server integrates with several key services:
βββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β Claude Desktop β
βββββββββββββ¬ββββββββββββββββββ¬ββββββββββββ¬ββββββββββββ
β β β
βΌ βΌ βΌ
βββββββββββββββββββββ βββββββββββββββ βββββββββββββββ
β NexusHub MCP β β Memory MCP β β GitHub MCP β
β Server β β Server β β Server β
βββββββββββ¬ββββββββββ βββββββββββββββ βββββββββββββββ
β β² β²
βΌ β β
βββββββββββββββββββ β β
β Vector Store β β β
βββββββββββββββββββ β β
β β β
βΌ β β
βββββββββββββββββββ βββββββββββββββ βββββββββββββββββββ
β Filesystem API β β Database APIβ β Docker API β
βββββββββββββββββββ βββββββββββββββ βββββββββββββββββββ
π οΈ Available Tools
NexusHub provides the following tools to Claude AI:
| Category | Tool | Description |
|----------|------|-------------|
| Filesystem | fs_list_files | List files in a directory |
| | fs_read_file | Read file contents |
| | fs_write_file | Write to a file |
| Database | db_execute_query | Execute a read-only SQL query |
| | db_list_tables | List all tables in the database |
| | db_describe_table | Get table schema |
| | db_insert_data | Insert data into a table |
| Docker | docker_list_containers | List Docker containers |
| | docker_start_container | Start a container |
| | docker_stop_container | Stop a container |
| | docker_get_container_logs | Get container logs |
| Search | serper_search | Perform web search via Serper API |
| Vector DB | ingest_docs | Ingest documents into the vector store |
| | vector_search | Search for similar documents |
π Getting Started
Prerequisites
- Docker and Docker Compose
- Node.js 18+
- Python 3.8+ (for vector embedding)
Installation
1. Clone the repository:
git clone https://github.com/webdevtodayjason/NexusHub.git
cd NexusHub
2. Set up environment variables:
cp .env.example .env
# Edit .env file with your API keys
3. Start the services:
docker-compose up -d
4. Access the dashboard:
Open http://localhost:8001/dashboard in your browser
Configuration for Claude Desktop
NexusHub supports two different integration methods with Claude Desktop:
1. HTTP Mode
This mode connects to NexusHub via HTTP endpoints. Add to your claude_desktop_config.json:
{
"mcpServers": {
"nexushub": {
"name": "NexusHub MCP Server",
"description": "Centralized MCP server with multiple capabilities",
"url": "http://localhost:8001/mcp",
"enabled": true
}
}
}
2. STDIO Mode (Recommended)
This mode uses direct process communication for better performance and reliability. It requires the NexusHub code to be available locally:
1. Prepare the wrapper scripts:
# Make the wrapper script executable
chmod +x /path/to/nexushub/nexushub-mcp.sh
chmod +x /path/to/nexushub/src/mcp/adapters/stdio-wrapper.js
chmod +x /path/to/nexushub/src/mcp/stdio-adapter.js
2. Add to your claude_desktop_config.json:
{
"mcpServers": {
"nexushub": {
"name": "NexusHub MCP Server",
"description": "Centralized MCP server with multiple capabilities",
"command": "/path/to/nexushub/nexushub-mcp.sh",
"enabled": true
},
"brave-search": {
"name": "Brave Search MCP",
"description": "Provides web search capabilities via Brave Search API.",
"command": "docker",
"args": ["exec", "-i", "mcp_brave_search", "node", "dist/index.js"],
"enabled": true
},
"github": {
"name": "GitHub MCP",
"description": "Provides GitHub repository interaction.",
"command": "docker",
"args": ["exec", "-i", "mcp_github", "node", "dist/index.js"],
"enabled": true
},
"memory": {
"name": "Memory MCP",
"description": "Persistent knowledge graph memory.",
"command": "docker",
"args": ["exec", "-i", "mcp_memory", "node", "dist/index.js"],
"enabled": true
}
}
}
Note: Replace /path/to/nexushub with your actual NexusHub installation path.
Docker Containers
Make sure all Docker containers are running before starting Claude Desktop:
cd /path/to/nexushub
docker-compose up -d
This will start all required MCP servers:
- NexusHub (primary server with filesystem, database, Docker tools)
- Brave Search MCP (web search capabilities)
- GitHub MCP (repository interaction)
- Memory MCP (knowledge graph/persistent memory)
π‘ Example Usage
Here are some examples of how to use NexusHub with Claude AI:
File Operations
Please use fs_list_files to show me all the JavaScript files in the src directory,
then read the content of any interesting files you find.
Database Query
Can you use db_execute_query to find all users in the database who joined
in the last month and have the "admin" role?
Vector Search
Please use vector_search to find documentation related to "authentication flow"
and summarize the key points.
Combined Tools
First, use brave_web_search to find the latest best practices for React error boundaries.
Then, use fs_read_file to examine our current implementation in ErrorBoundary.jsx,
and suggest improvements based on what you found.
π§© How to Add Your Own Tools
You can extend NexusHub with custom tools by adding new tool definitions to the /src/mcp/tools/ directory:
// src/mcp/tools/my-custom-tools.js
export function getToolDefinitions() {
return {
my_custom_tool: {
name: 'my_custom_tool',
description: 'Description of what the tool does.',
inputSchema: {
type: 'object',
properties: {
param1: {
type: 'string',
description: 'Description of parameter 1'
}
// Add more parameters as needed
},
required: ['param1']
}
}
};
}
// Tool implementation
export async function myCustomTool(param1) {
// Your custom tool logic here
return { result: 'Success!' };
}
Then add the tool to /src/mcp/tools/index.js and /src/mcp/tools/handler.js.
π Local Development
1. Install dependencies:
npm install
2. Run in development mode:
npm run dev
3. For stdio mode testing:
npm run stdio
π§ͺ Running Tests
```bash
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.



