AI Path Finder
Free consultation

Lesson · Hinglish · ⏱ 6 min padhne mein

🔗 MCP: AI tools ke liye universal plug

EnglishHinglish

🏠 Asli zindagi ki misaal

USB-C se pehle har phone ka apna charger hota tha. Ab ek cable lagbhag sab cheezon mein lag jaati hai. Model Context Protocol (MCP) AI ke liye yahi karta hai: apne tool ya data ka connector ek baar banao, aur bahut saari AI apps aur agents usmein plug ho sakte hain.

💡 Aasaan shabdon mein

MCP AI applications ko tools aur data se jodne ka ek open standard hai. MCP server tools (actions), resources (padhne ka data) aur prompts ek standard format mein deta hai. MCP client, jaise AI assistant ya agent, unhe dhoondhkar use karta hai. Isliye har AI app ke liye custom integration code likhne ki jagah, aap ek MCP server likhte ho, jaise company ke ticketing system ke liye. Engineers ke liye MCP server banaana ek practical, demand wali skill hai, jo dikhaati hai ki aap tool use, auth aur security samajhte ho.

🪜 Step by step

  1. MCP server ek system (tickets, files, database) ko wrap karta hai aur tools, resources aur prompts standard format mein deta hai.
  2. MCP client ek AI app (assistant, IDE, aapka apna agent) ke andar rehta hai aur ek ya zyaada servers se judta hai.
  3. Client server ke tools ki list leta hai aur model ko deta hai, phir model unhe call kar sakta hai.
  4. Calls client se server tak jaati hain, jo apne authentication aur permissions ke saath unhe chalata hai.
  5. Ek baar banao, har jagah use karo: wahi server har MCP compatible app ke saath chalta hai.

🔑 Zaroori shabd

MCP server
Program jo tools aur data ko MCP format mein AI apps ko deta hai.
MCP client
AI app ya agent jo MCP servers se judta hai.
Resource
Read-only data jo server share karta hai, jaise file ya record.

👀 Example dekho

python
from mcp.server.fastmcp import FastMCP
mcp = FastMCP("tickets")

@mcp.tool()
def open_tickets(team: str) -> list[dict]:
    """List open support tickets for a team."""
    return db.query("SELECT id, title FROM tickets WHERE team=? AND open=1", team)

mcp.run()

Ek chhota server, aur ab koi bhi MCP-compatible assistant tickets list kar sakta hai.

⚠️ Aam galtiyaan jinse bachna hai

  • Anjaan sources se MCP servers install karna. Server data padh sakta hai aur actions le sakta hai; ise kisi bhi install kiye software ki tarah dekho.
  • Default mein write ya delete tools dena. Read only se shuru karo aur powerful tools dhyaan se jodo.

✅ Jaldi se check karo

MCP ka main faayda kya hai?

  1. Models ko smart banata hai
  2. Tool connector ek baar banao aur AI apps mein reuse karo
  3. Databases ko replace karta hai
Jawaab dekho

B. Tool connector ek baar banao aur AI apps mein reuse karo: MCP standard banata hai ki tools kaise expose hon, jaise AI integrations ke liye USB-C.

🙋 Seekhne wale yeh poochhte hain

Kya MCP sirf ek company ke models ke liye hai?

Nahi. Yeh open protocol hai jise kai AI apps aur model providers support karte hain.

MCP ya normal tool calling?

Dono saath kaam karte hain. Tool calling woh tarika hai jisse model action maangta hai; MCP un tools ko apps ke beech pack aur share karne ka standard tarika hai.

📌 Yaad rakho

MCP = AI apps aur aapke tools/data ke beech ek standard plug.

📖 Is lesson ke shabd (AI dictionary)

Poora course, aapke roadmap ke hisaab se

Progress tracking, quick checks aur step-by-step order ke saath. Yeh lesson in courses ka hissa hai:

LLM / Agent Engineer

Course mein seekho