MCP in Practice
MCP Is More Than Just "Calling Tools"
Most people understand MCP as simply "letting AI call external tools." But the MCP protocol is bidirectional: Alice can call other tools, and external clients like Cursor can also call Alice's capabilities.
Bidirectional Flow Diagram
Calendar
Email
Database
Web
→
←
Alice calls tools
A
Alice
Both Client and Server
←
→
External calls Alice
Cursor
Claude Desktop
⚙️Automation Script
Two Roles, One Protocol
As Client (Consumer)
Alice connects to external MCP services and calls tool capabilities provided by others:
- Calendar MCP — view and create events
- Email MCP — read and write emails
- Database MCP — query data
- Browser MCP — access web pages
Key: Alice proactively initiates requests to acquire external capabilities.
As Server (Provider)
Alice exposes its own capabilities via the MCP protocol to be called externally:
- Cursor — calls Alice to search notes while coding
- Claude Desktop — lets Claude send messages to Alice
- Automation scripts — scheduled triggers to make Alice execute tasks
Key: Alice passively receives requests and provides its own capabilities.
Why does bidirectionality matter?
One-way integration is just tool calling — no different from an API. Bidirectional integration means AI can not only use tools but also become a tool. When multiple AI Agents can call each other, an AI ecosystem naturally emerges.
One-way integration is just tool calling — no different from an API. Bidirectional integration means AI can not only use tools but also become a tool. When multiple AI Agents can call each other, an AI ecosystem naturally emerges.
MCP is not just about AI calling tools — it lets AI be called by other AI systems, and that is the power of an ecosystem. One protocol, bidirectional integration: the critical leap from tool to platform.