Scribd
Upload
654 views10 pages

Anthropic MCP Server

The Model Context Protocol (MCP) is an open standard launched by Anthropic in November 2024 that facilitates communication between AI clients and data servers, allowing real-time data access without custom code. MCP supports various applications including multi-step projects, real-world awareness, and enhanced customer support by enabling AI to interact with structured context. The document outlines setup requirements, hands-on tasks, and the potential uses of MCP in personal and collaborative AI environments.

Uploaded by

shubham jha
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
654 views10 pages

Anthropic MCP Server

The Model Context Protocol (MCP) is an open standard launched by Anthropic in November 2024 that facilitates communication between AI clients and data servers, allowing real-time data access without custom code. MCP supports various applications including multi-step projects, real-world awareness, and enhanced customer support by enabling AI to interact with structured context. The document outlines setup requirements, hands-on tasks, and the potential uses of MCP in personal and collaborative AI environments.

Uploaded by

shubham jha
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

MCP

A GUIDE ON

MODEL CONTEXT
PROTOCOL (MCP)
WORKING
HANDS-ON
APPLICATIONS
& MORE
MCP

WHAT IS MCP?
It is a powerful open standard that was launched by Claude’s
parent company, Anthropic, in November 2024.

Model: The LLM (e.g., Claude, GPT-4) generates responses.

Context: Additional input (documents, PDFs, prompts,


databases) for meaningful replies.

Protocol: Rules enabling the model to access and use


structured context.
MCP

WHAT CAN WE DO
WITH MCP?
MCP streamlines communication between the AI clients and
data servers, enabling the system to access real-time data from
various sources without custom code or manual uploads.

Client: The interface for LLM responses (e.g., Claude's app, IDE-
like cursor, custom chatbots, etc).
Server: The data source storing context (e.g., Google Drive,
GitHub, database, local files, etc ).
MCP

GROWING POPULARITY
OF MCP
Clients Servers
MCP

MCP: HANDS ON
Tasks:

1 Create a Custom Greeting

2 Count the number of files on our desktop

3 Save a chat with Claude to our local system

4 Ask Claude questions using a local PDF


document

To perform these 4 tasks, we are working with :

Client: Claude Desktop App

Server: Local files like documents and PDFs that are available
on the local system
MCP

MCP: SET UP REQUIRMENTS


Step 1: Create a Python Environment
With Conda type, conda create ~n mcp python==3.11

Step 2 : Install the MCP Library


pip install mcp

Step 3 : Download Claude’s Desktop App

Download the app from https://claude.ai/download


Works with free plan as well but may hit rate limits while
experimenting

Step 4 : Set Up the Configuration File


On the Desktop app, go to Claude Settings.
Then go to the Developer Section.
Click on “Edit Configuration”
Look for or create “claude_desktop_config.json”
MCP

TASK 1 : CUSTOM GREETING


Step 1: Write the code in a new Python file
Name the file: greeter.py

Code:

An MCP server listens for greetings ("Hi", "Hey", "Hello") and responds with a
personalized message.
The server is created simply using FastMCP and the tool is added using
@mcp.tool() decorator.
Outputs can differ since the LLM decides the final response.
MCP

TASK 1 : CUSTOM GREETING


Step 2: Update the configuration file
Name of our server
Location of the Python program
Location of the Python file that has to be executed.

Step 3: Restart Your Claude Desktop App


Any error in the code gets highlighted on this step.
If no error is found, the tool gets registered within the desktop app
Perform this step every time the code is changed or updated..
MCP

TASK 1 : CUSTOM GREETING


Step 4: Test the tool
Just prompt a “Hi” to Claude and when asked, permit the LLM to use the tool.

Output:
MCP

MCP: APPLICATIONS
Multi-Step Projects: MCP enables AI to coordinate tasks across platforms,
like event planning, without complex integrations.

Real-World Awareness: MCP helps AI interact with smart environments and IoT
devices, making it more proactive.

Collaborating Agents: MCP allows multiple AI agents to collaborate and share


information without direct integrations.

Personal AI Assistants: MCP enables personalized assistants to securely access


personal data without exposing it to third parties.

Enhanced Customer Support: With MCP, AI provides context-aware customer


service by accessing real-time data, improving efficiency.

You might also like