MCP & Integrations
MCP -- The USB Standard for AI
Imagine every AI application had to invent its own proprietary way to communicate with external data sources and tools. That was exactly the reality until late 2025. Then came the Model Context Protocol (MCP).
MCP is an open protocol that standardizes how AI models interact with the outside world. The USB analogy is apt: before USB, every device had its own connector. USB created a universal standard -- and that's exactly what MCP does for AI integrations.
The Origin Story
MCP was released by Anthropic in November 2024 as an open-source protocol. In December 2025, it was handed over to the Agentic AI Foundation (AAIF) under the Linux Foundation -- a clear sign that the entire industry supports this approach.
Why Does This Matter?
Without MCP, every AI provider had to build custom integrations:
- ChatGPT needed its own plugins
- Claude needed its own tool definitions
- Every company had to implement integrations multiple times
With MCP, you build a server once -- and it works with every MCP-enabled AI application.
iMCP by the Numbers (as of 2026)
Over 10,000 MCP servers are publicly available. The protocol is supported by ChatGPT, Claude, Gemini, Cursor, VS Code Copilot, and many other tools. Adoption has been faster than any previous AI integration.
Before vs. After
| Aspect | Without MCP | With MCP |
|---|---|---|
| Building integrations | Separately for each AI tool | Build once, use everywhere |
| Protocol | Proprietary, varying | JSON-RPC 2.0, standardized |
| Maintenance | Maintain N integrations | Maintain 1 server |
| Security | Varies by provider | Unified auth model |
| Community | Fragmented | Central ecosystem |
What You'll Learn in This Module
In the following sections, you will:
- Understand the MCP architecture -- Host, Client, Transport, Server, and how they work together
- Learn the three primitives -- Tools, Resources, and Prompts as building blocks
- Build your own MCP server -- Step by step in TypeScript
- Analyze enterprise use cases -- How companies use MCP for Jira, Slack, and more
*Learning Objective
After this module, you will be able to explain the MCP architecture, distinguish between the three primitives, and implement your own MCP server. You'll be operating at Bloom's levels of "Apply" and "Analyze."
Why is MCP often called the 'USB standard for AI'?
Let's get started -- with a look at the architecture behind MCP.
Reflect
MCP standardizes communication between AI and external services -- just as USB standardized hardware connections. An open standard supported by all major AI providers. Let us start with a look at the architecture.