MCP Builder App
C# · .NET MAUI · SQLite
A .NET MAUI desktop application that generates Model Context Protocol (MCP) servers from documentation site URLs, enabling AI agents to interact with sites that lack native MCP support. Implements a hybrid C# scraping pipeline that attempts fast HTML parsing via HtmlAgilityPack with automatic fallback to Selenium for JavaScript-rendered pages, balancing speed and coverage.
Overview / The Problem
MCP Builder App is a cross-platform desktop tool built with .NET MAUI that empowers developers and AI practitioners to generate Model Context Protocol (MCP) servers from any documentation website. Many popular libraries and frameworks lack native MCP support, making it difficult for AI agents like Claude or Copilot to interact with their documentation programmatically. MCP Builder bridges that gap by scraping documentation sites and producing a fully functional MCP server that AI agents can query.
How It Works
The application uses a hybrid scraping pipeline written entirely in C#. It first attempts fast HTML parsing using HtmlAgilityPack to extract content from static pages. If JavaScript-rendered content is detected, the pipeline automatically falls back to Selenium WebDriver for full browser-based scraping. This dual approach balances speed with coverage, ensuring that even modern single-page application documentation sites are fully captured. All scraped data is stored locally in an SQLite database for offline access and fast server generation.
Architecture & Design
The project is built on .NET MAUI for the desktop UI, leveraging the MVVM pattern for clean separation of concerns. The scraping engine uses HtmlAgilityPack and Selenium WebDriver, with a strategy pattern to switch between them seamlessly. Data persistence is handled through SQLite with a lightweight repository layer. The generated MCP servers follow the official Model Context Protocol specification, ensuring compatibility with any MCP-compliant AI agent.