VectorMCP
A Ruby gem for building Model Context Protocol (MCP) servers to expose tools, resources, and prompts to LLM clients.
VectorMCP
VectorMCP is a Ruby gem implementing the Model Context Protocol (MCP) server-side specification. It provides a framework for creating MCP servers that expose tools, resources, prompts, and roots to LLM clients.
Why VectorMCP?
- 🛡️ Security-First: Built-in input validation and schema checking prevent injection attacks
- ⚡ Production-Ready: Robust error handling, comprehensive test suite, and proven reliability
- 🔌 Multiple Transports: stdio for CLI tools, SSE for web applications
- 📦 Zero Configuration: Works out of the box with sensible defaults
- 🔄 Fully Compatible: Implements the complete MCP specification
Quick Start
gem install vector_mcp
require 'vector_mcp'
# Create a server
server = VectorMCP.new(name: 'MyApp', version: '1.0.0')
# Add a tool
server.register_tool(
name: 'greet',
description: 'Says hello to someone',
input_schema: {
type: 'object',
properties: { name: { type: 'string' } },
required: ['name']
}
) { |args| "Hello, #{args['name']}!" }
# Start the server
server.run # Uses stdio transport by default
That's it! Your MCP server is ready to connect with Claude Desktop, custom clients, or any MCP-compatible application.
Transport Options
Command Line Tools (stdio)
Perfect for desktop applications and process-based integrations:
server.run # Default: stdio transport
Web Applications (HTTP + SSE)
Ideal for web apps and browser-based clients:
server.run(transport: :sse, port: 8080)
Connect via Server-Sent Events at http://localhost:8080/sse
Core Features
Tools (Functions)
Expose functions that LLMs can call:
server.register_tool(
name: 'calculate',
description: 'Performs basic math',
input_schema: {
type: 'object',
properties: {
operation: { type: 'string', enum: ['add', 'subtract', 'multiply'] },
a: { type: 'number' },
b: { type: 'number' }
},
required: ['operation', 'a', 'b']
}
) do |args|
case args['operation']
when 'add' then args['a'] + args['b']
when 'subtract' then args['a'] - args['b']
when 'multiply' then args['a'] * args['b']
end
end
Resources (Data Sources)
Provide data that LLMs can read:
server.register_resource(
uri: 'file://config.json',
name: 'App Configuration',
description: 'Current application settings'
) { File.read('config.json') }
Prompts (Templates)
Create reusable prompt templates:
server.register_prompt(
name: 'code_review',
description: 'Reviews code for best practices',
arguments: [
{ name: 'language', description: 'Programming language', required: true },
{ name: 'code', description: 'Code to review', required: true }
]
) do |args|
{
messages: [{
role: 'user',
content: {
type: 'text',
text: "Review this #{args['language']} code:\n\n#{args['code']}"
}
}]
}
end
Security Features
VectorMCP provides comprehensive, opt-in security for production applications:
Built-in Input Validation
All inputs are automatically validated against your schemas:
# This tool is protected against invalid inputs
server.register_tool(
name: 'process_user',
input_schema: {
type: 'object',
properties: {
email: { type: 'string', format: 'email' },
age: { type: 'integer', minimum: 0, maximum: 150 }
},
required: ['email']
}
) { |args| "Processing #{args['email']}" }
# Invalid inputs are automatically rejected:
# ❌ { email: "not-an-email" } -> Validation error
# ❌ { age: -5 } -> Missing required field
# ✅ { email: "[email protected]" } -> Passes validation
Authentication & Authorization
Secure your MCP server with flexible authentication strategies:
# API Key Authentication
server.enable_authentication!(
strategy: :api_key,
keys: ["your-secret-key", "another-key"]
)
# JWT Token Authentication
server.enable_authentication!(
strategy: :jwt,
secret: ENV["JWT_SECRET"]
)
# Custom Authentication Logic
server.enable_authentication!(strategy: :custom) do |request|
api_key = request[:headers]["X-API-Key"]
User.find_by(api_key: api_key) ? { user_id: user.id } : false
end
Fine-Grained Authorization
Control access to tools, resources, and prompts:
server.enable_authorization! do
# Tool-level access control
authorize_tools do |user, action, tool|
case user[:role]
when "admin" then true
when "user" then !tool.name.start_with?("admin_")
else false
end
end
# Resource-level permissions
authorize_resources do |user, action, resource|
user[:tenant_id] == resource.tenant_id
end
end
Transport Security
Security works seamlessly across all transport layers:
- Stdio: Header simulation for desktop applications
- SSE: Full HTTP header and query parameter support
- Request Pipeline: Automatic authentication and authorization checking
Our comprehensive security documentation covers authentication strategies, authorization policies, session management, and real-world examples.
Real-World Examples
File System Server
server.register_tool(
name: 'read_file',
description: 'Reads a text file',
input_schema: {
type: 'object',
properties: { path: { type: 'string' } },
required: ['path']
}
) { |args| File.read(args['path']) }
Database Query Tool
server.register_tool(
name: 'search_users',
description: 'Searches users by name',
input_schema: {
type: 'object',
properties: {
query: { type: 'string', minLength: 1 },
limit: { type: 'integer', minimum: 1, maximum: 100 }
},
required: ['query']
}
) do |args|
User.where('name ILIKE ?', "%#{args['query']}%")
.limit(args['limit'] || 10)
.to_json
end
API Integration
server.register_tool(
name: 'get_weather',
description: 'Gets current weather for a city',
input_schema: {
type: 'object',
properties: { city: { type: 'string' } },
required: ['city']
}
) do |args|
response = HTTP.get("https://api.weather.com/current", params: { city: args['city'] })
response.parse
end
Advanced Usage
Filesystem Roots & Security
Define secure filesystem boundaries:
# Register allowed directories
server.register_root_from_path('./src', name: 'Source Code')
server.register_root_from_path('./docs', name: 'Documentation')
# Tools can safely operate within these bounds
server.register_tool(
name: 'list_files',
input_schema: {
type: 'object',
properties: { root_uri: { type: 'string' } },
required: ['root_uri']
}
) do |args|
root = server.roots[args['root_uri']]
raise 'Invalid root' unless root
Dir.entries(root.path).reject { |f| f.start_with?('.') }
end
LLM Sampling (Server → Client)
Make requests to the connected LLM:
server.register_tool(
name: 'generate_summary',
input_schema: {
type: 'object',
properties: { text: { type: 'string' } },
required: ['text']
}
) do |args, session|
result = session.sample(
messages: [{
role: 'user',
content: { type: 'text', text: "Summarize: #{args['text']}" }
}],
max_tokens: 100
)
result.text_content
end
Custom Error Handling
Use proper MCP error types:
server.register_tool(name: 'risky_operation') do |args|
if args['dangerous']
raise VectorMCP::InvalidParamsError.new('Dangerous operation not allowed')
end
begin
perform_operation(args)
rescue SomeError => e
raise VectorMCP::InternalError.new('Operation failed')
end
end
Session Information
Access client context:
server.register_tool(name: 'client_info') do |args, session|
{
client: session.client_info&.dig('name'),
capabilities: session.client_capabilities,
initialized: session.initialized?
}
end
Integration Examples
Claude Desktop
Add to your Claude Desktop configuration:
{
"mcpServers": {
"my-ruby-server": {
"command": "ruby",
"args": ["path/to/my_server.rb"]
}
}
}
Web Applications
// Connect to SSE endpoint
const eventSource = new EventSource('http://localhost:8080/sse');
eventSource.addEventListener('endpoint', (event) => {
const { uri } = JSON.parse(event.data);
// Send MCP requests
fetch(uri, {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({
jsonrpc: '2.0',
id: 1,
method: 'tools/call',
params: { name: 'greet', arguments: { name: 'World' } }
})
});
});
Why Choose VectorMCP?
🏆 Battle-Tested: Used in production applications serving thousands of requests
⚡ Performance: Optimized for low latency and high throughput
🛡️ Secure by Default: Comprehensive input validation prevents common attacks
📖 Well-Documented: Extensive examples and clear API documentation
🔧 Extensible: Easy to customize and extend for your specific needs
🤝 Community: Active development and responsive maintainer
Examples & Resources
- Examples Directory - Complete working examples
- File System MCP - Real-world implementation
- MCP Specification - Official protocol documentation
Installation & Setup
gem install vector_mcp
# Or in your Gemfile
gem 'vector_mcp'
Contributing
Bug reports and pull requests welcome on GitHub.
License
Available as open source under the MIT License.
相關伺服器
Scout Monitoring MCP
贊助Put performance and error data directly in the hands of your AI assistant.
Alpha Vantage MCP Server
贊助Access financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
Taeks MCP Server
An example MCP server designed for deployment on Cloudflare Workers, operating without authentication.
MCP-Insomnia
An MCP server for AI agents to create and manage API collections in Insomnia-compatible format.
SSH Server MCP
An MCP server that provides SSH-based remote management tools, acting as proxy
MCP SBOM Server
Performs a Trivy scan to produce a Software Bill of Materials (SBOM) in CycloneDX format.
MCP LSP Go
An MCP server that connects AI assistants to Go's Language Server Protocol (LSP) for advanced code analysis.
vnsh
Ephemeral encrypted file sharing for AI. Client-side AES-256 encryption, 24h auto-vaporization.
MCP Lab
A development environment for building and testing custom MCP servers with AI and VS Code integration.
Figma MCP Server
Enables AI assistants to interact with Figma via WebSocket for reading data and design analysis.
MediaWiki
Interact with MediaWiki installations through the MediaWiki API as a bot user.
TeamCity MCP Server
MCP server for JetBrains TeamCity with 87 tools for builds, tests, agents, and CI/CD pipeline management.