MCP Contd
MCP Architecture
- Overview

- MCP Server:
- This is server side of mcp which works with external datasources (Google Drive, Slack, Github…………….)
- MCP Host:
- This is an IDE/LLM Chat Interface/Python code
- Host creates one mcp client per mcp server
-
MCP Client:
- Interacts with MCP Server
-
- Stdio transport:
- MCP Client and MCP Server on same machine
- This uses Remote Procedure Calls
- Streamable HTTP transport
- MCP Client and MCP Server on different Machines
- It uses HTTPS streams

- Stdio transport:
-
Overview

MCP SDKs
- Client SDKS

- Server SDK

Claude Desktop
- Download Claude Desktop Refer Here
UV
- Refer Here for UV
- Installation Refer Here
- Open Terminal
pip install uv
Here is a concise cheatsheet for uv, the fast Python package and project manager:
uv Cheatsheet
Project Setup & Management
- Initialize a new project:
uv init[1][3] - Add a package:
uv addoruv add @[1][3] - Remove a package:
uv remove[1][3] - Lock dependencies:
uv lock[1][4] - Install all dependencies:
uv install[1] - Update all dependencies:
uv updateoruv lock --upgrade[1][4] - List outdated dependencies:
uv outdated[1] - List all installed dependencies:
uv list[1] - Clean unused dependencies:
uv clean[1] - View dependency tree:
uv tree[3]
Virtual Environments
- Create a virtual environment:
uv env createoruv sync(creates env if needed)[1][4] - Activate a virtual environment:
uv env activate[1] - Deactivate a virtual environment:
uv env deactivate[1]
Python Version Management
- Install a specific Python version:
uv python install[1][3] - Display current Python version:
uv python --version[1] - List available Python versions:
uv python list[3]
Running Scripts & Tools
- Run a Python script:
uv run[1][3] - Run a tool:
uvx[1][3] - Install a command-line tool:
uv tool install[1][3]
Miscellaneous
- Publish a package:
uv publish[1] - Search for available tools:
uv tool search[1] - List installed tools:
uv tool list[1] - Help:
uv help[1][5] - Upgrade uv:
uv upgradeoruv self update[1][5]
Cache Management
- Clean cache:
uv cache clean[5] - Prune cache:
uv cache prune[5] - Show cache directory:
uv cache dir[5]
This covers the most common and essential commands for working with uv in Python projects[1][3][5].
