MCP Multi-Feature Server

This project is a Model Context Protocol (MCP) server implemented using Node.js and TypeScript, following the official MCP TypeScript SDK.
It provides multiple capabilities including weather data, system utilities, Azure integration, and AI-powered image generation and editing.

Features

• Current Weather: Get the current weather for any city
• Weather Forecast: Get a 3-day weather forecast for any city
• City Search: Find cities by name
• CLI Commands: Execute safe, whitelisted CLI commands
• Directory Listing: List files and directories
• System Info: Get basic system information
• Azure Integration: Manage Azure subscriptions and resource groups with MCP elicitation
• Azure Storage: Manage Azure Storage blobs and containers
• Image Generation: Generate images using AI models (gpt-image-1, flux.1-kontext-pro)
• Image Editing: Edit existing images with AI-powered modifications and enhancements

Getting Started

Prerequisites

• Node.js v18.x or higher

Install dependencies

npm install

Configure Environment

1. Copy the sample environment file:

cp .env.sample .env

2. Edit .env file to configure your image API endpoints:

GENERATE_IMAGE_API_URL=http://127.0.0.1:8000/v1/images/generations
EDIT_IMAGE_API_URL=http://127.0.0.1:8000/v1/images/edits

Build the project

npx tsc

Run the server (development)

npx ts-node src/server.ts

Run the server (compiled)

npm run start

Project Structure

• src/server.ts: Main MCP server implementation
• package.json: Project configuration and dependencies
• tsconfig.json: TypeScript configuration
• .env: Environment configuration (create from .env.sample)
• .env.sample: Environment configuration template
• .gitignore: Git ignore rules

Sample Prompts

You can use these prompts with an MCP-compatible client or extension:

• Get current weather

  • currentWeather: city = Seattle
  • What is the current weather in Paris?
  • weather in Phoenix

• Get weather forecast

  • forecast: city = New York
  • weather forecast for Seattle
  • 3-day forecast for London

• Search for a city

  • searchCity: query = Washington
  • Find cities named "Springfield"

• Execute CLI commands

  • executeCommand: command = "git status"
  • executeCommand: command = "ls -la"
  • executeCommand: command = "npm --version"

• List directory contents

  • listDirectory: path = "."
  • listDirectory: path = "src"

• Get system information

  • getSystemInfo

• Azure subscription management

  • listAzureSubscriptions
  • getCurrentAzureSubscription

• Azure resource groups (with MCP elicitation)

  • listAzureResourceGroups (uses VS Code command palette for subscription selection)

• Image generation

  • generateImages: prompt = "A futuristic cityscape at sunset"
  • generateImages: prompt = "A cute robot", model = "flux.1-kontext-pro", size = "512x512", quality = "hd"
  • generateImages: prompt = "Mountain landscape", model = "gpt-image-1", n = 2

• Image editing

  • editImages: prompt = "Add a rainbow in the sky", image = "<base64-data>"
  • editImages: prompt = "Change the car color to red", image = "<base64-data>", model = "gpt-image-1"
  • editImages: prompt = "Remove the background", image = "<base64-data>", quality = "hd"

Configuration

Environment Variables

Create a .env file in the project root (copy from .env.sample):

# Image Generation API Configuration
GENERATE_IMAGE_API_URL=http://127.0.0.1:8000/v1/images/generations

# Image Editing API Configuration  
EDIT_IMAGE_API_URL=http://127.0.0.1:8000/v1/images/edits

Note: You can configure separate endpoints for generation and editing to work with different services.

Image Generation API Requirements

The image generation tools require running API servers with the following endpoints:

• POST /v1/images/generations - Generate images from text prompts (configured via GENERATE_IMAGE_API_URL)
• POST /v1/images/edits - Edit images with AI modifications (configured via EDIT_IMAGE_API_URL)

Supported Models:

• gpt-image-1 (Azure OpenAI DALL-E) - Default model
• flux.1-kontext-pro (Flux model) - High-quality artistic generation

Model-Specific Features:

• gpt-image-1: Quality options (standard, hd), supports both generation and editing
• flux.1-kontext-pro: Standard quality only, optimized for creative content

Supported Sizes: 1024x1024 (default), 512x512, 256x256, and other standard dimensions Quality Options:

• gpt-image-1: standard, hd
• flux.1-kontext-pro: standard

MCP Elicitation Features ⭐

This server demonstrates proper implementation of the MCP Elicitation specification with VS Code integration:

listAzureResourceGroups Tool

• Interactive Selection: When called without a subscriptionId parameter, triggers the VS Code MCP extension's command palette
• Native Integration: Uses the official MCP elicitation protocol (elicitation/create JSON-RPC request)
• Rich UI: Shows subscription names with "(ACTIVE)" indicator in the selection dropdown
• Enum Schema: Provides structured choices with display names for better UX

How it works:

1. Tool is called without subscriptionId parameter
2. Server sends elicitation/create request to client
3. VS Code MCP extension shows command palette with subscription options
4. User selects subscription from dropdown
5. Tool continues with selected subscription ID

This showcases the proper way to implement interactive, user-driven parameter selection in MCP servers.

References

• Model Context Protocol TypeScript SDK