CLAUDE Desktop w Secure MCP AI Agents (Anthropic)

How to Get started with the Model Context Protocol (MCP) by Anthropic.
My deep dive into the new protocol implementation by Anthropic for a secure AI to Data connection via the new MCP Server architecture.

We will code a MCP server together (w MCP Python SDK) and examine the new architecture idea, that should provide a secure protocol from Claude LLMs to their tools, data sources and external APIs, as defined by ‪@anthropic-ai‬

All rights w authors:
https://github.com/modelcontextprotoc...
https://modelcontextprotocol.io/docs/...
https://modelcontextprotocol.io/docs/...


00:00 New Model Context Protocol MCP
03:46 LangChain vs Secure MCP Server
08:22 Focus on Security AI Agents LLM
09:25 Client Server Protocol Architecture MCP
13:07 Python MCP SDK pre-build server
16:19 Internet Search MCP Server for CLAUDE
17:40 SQLite MCP Server w MCP Protocol
20:02 MCP architecture Host, Client, Server
25:10 Let us build our first MCP Server
27:54 Protocol layer of MCP
28:20 MCP Resources
29:20 MCP Prompts
30:30 MCP Tools
34:54 MCP Sampling
36:00 Outlook



#ai
#anthropic
#programming
#aiagents
#airesearch
#coding
#protocol