The Model Context Protocol (MCP) changes everything. It gives LLMs like Claude and ChatGPT the ability to use real tools through APIs instead of clunky GUIs.
In this video, I’ll break down how MCP works, why it’s so powerful, and how you can run it using Docker containers. I’ll even show you how to build your own custom MCP servers so you can connect AI to literally ANYTHING: Obsidian, Brave, Kali Linux, and more. Grab your coffee.
Let’s go!
00:00 - Intro
00:40 - Problem: LLMs Suck at Accessing Code
02:40 - What is MCP?
04:06 - Running Your First MCP Server
07:32 - Client 1: Claude Desktop
10:12 - Client 2: LM Studio (Local Models)
11:48 - Client 3: Cursor (IDE)
12:31 - Build YOUR OWN MCP Server
19:05 - Building More Complex MCPs
23:19 - Kali Linux HACKING MCP
28:00 - How does this actually work?
32:51 - MCP Gateway, Explained
35:53 - Running an MCP Gateway with n8n