The MCP Server Builder scaffolds complete MCP server projects for VS Code and Cursor using a single create_mcp_server tool.
Key Capabilities:
Generate MCP Server Projects: Create fully-configured projects with proper structure, TypeScript setup, MCP SDK integration, and development workflow with hot reload
Intelligent File Analysis: Automatically detect and include appropriate MCP capabilities (resources, tools, prompts) by analyzing files (JSON, YAML, XML, CSV, TXT, MD, config files)
Flexible Configuration: Customize project name (required), description, author, output path, and choose between current directory or new subdirectory creation
Auto-Detect Resources: Automatically enable resources capability when data files are detected, or manually include using
includeResourcesparameterProduction-Ready Setup: Generated projects include TypeScript with strict configuration, build tools, development scripts (build, dev, start), and example implementations to help you get started quickly
Provides access to Ferengi Rules of Acquisition data from Star Trek, used as an example dataset in the tutorial for building MCP servers that serve structured data.
MCP Server Builder
A minimalist MCP server that scaffolds basic MCP server projects for VS Code and Cursor.
For Users
Installation
MCP Configuration
Add this to your editor's MCP settings configuration files:
VS Code:
Cursor:
Note: Replace /absolute/path/to/MCP-Builder with the actual absolute path where you cloned this repository.
Usage
Once configured, use the create_mcp_server tool to generate new MCP server projects:
name (required): Project name (kebab-case recommended)
description (optional): Project description
author (optional): Author name
outputPath (optional): Target directory (defaults to current working directory)
includeResources (optional): Force include resources capability (default: false)
analyzeFiles (optional): File paths to analyze for auto-determining capabilities
createSubdirectory (optional): Create project in new subdirectory (default: false)
Intelligent File Analysis
The builder can analyze your files to automatically determine what capabilities to include:
Auto-detects Resources for:
Data files:
.json,.yaml,.xml,.csv,.txt,.mdConfig files:
.config,.env, files named "config"Directories containing data files
Files with structured content (JSON/YAML patterns)
Directory Behavior
Default (Current Directory):
→ Creates files directly in current directory
Subdirectory Mode:
→ Creates ./my-server/ subdirectory with files
Tutorial: Your First MCP Server with Example Data
This project includes a fun example JSON dataset to help you create your first MCP server: Ferengi Rules of Acquisition from Star Trek! This is perfect for learning how to build an MCP server that serves data.
Once you get the MCP Server Builder set up, you can use the create_mcp_server tool to generate a server that provides access to this dataset - or just use a prompt. The builder will analyze the JSON file and automatically include the Resources capability, making it easy to serve the data to your AI assistant.
What's Included
The docs/ferengi-rules-of-acquisition.json file contains 200+ Rules of Acquisition—the sacred commercial guidelines of the Ferengi species. Each rule includes:
Rule number
The rule text
Source (episode, novel, or game)
Create Your First Server
Step 1: Copy the example data to a new folder
Step 2: Ask your AI assistant to build the server
In VS Code/Cursor, simply tell your AI assistant:
"Create an MCP server in the ferengi-rules-server folder that provides access to Ferengi Rules of Acquisition. Analyze the ferengi-rules-of-acquisition.json file in that directory."
That's it! Your AI assistant will use the create_mcp_server tool to:
Detect the JSON data file
Automatically include the Resources capability
Generate a complete MCP server project with all the scaffolding
Set up TypeScript, build tools, and development workflow
Ideas for Your Server
Once you have the basic server running, you could add:
Tools:
get_rule_by_number- Fetch a specific rulesearch_rules- Search by keywordrandom_rule- Get a random rule for inspirationrules_by_source- Filter by episode or book
Resources:
ferengi://rules/all- All rules as a resourceferengi://rules/{number}- Individual rule by numberferengi://rules/random- Random rule
Prompts:
Help users apply Ferengi wisdom to their business decisions
Generate "Ferengi-style" advice for scenarios
Example prompts to add these features:
"Add a tool called
get_rule_by_numberto the Ferengi server that takes a rule number and returns that specific rule from the JSON file."
"Add a
search_rulestool that searches through all the rules and returns any that contain the search keyword in the rule text."
"Create a
random_ruletool that returns a random Ferengi Rule of Acquisition for inspiration."
This example demonstrates how any JSON data source can become an MCP server that AI assistants can query and use!
Generated Project Features
TypeScript with strict configuration
MCP SDK integration with best practices
Example tool implementation
Optional resources capability (auto-detected or manual)
Development workflow with hot reload
Comprehensive build setup
For Developers
Development Setup
Scripts
npm run build- Compile TypeScriptnpm run dev- Development mode with hot reloadnpm start- Run compiled server
Project Structure
Contributing
The generated projects include everything needed to start building MCP servers immediately. When making changes to the builder:
Test with
npm run devfor hot reloadBuild with
npm run buildTest the generated servers to ensure they work correctly