MCP & A2A - Model Context Protocol & Agent to Agent Protocol

Kartik Marwah, Teaching AI, Agents, MCP, A2A

Tenha acesso ilimitado a todos os cursos

Oferecidos por líderes do setor e profissionais do mercado

Os temas incluem ilustração, design, fotografia e muito mais

Tenha acesso ilimitado a todos os cursos

Oferecidos por líderes do setor e profissionais do mercado

Os temas incluem ilustração, design, fotografia e muito mais

Aulas neste curso

- 1.
  
  Course Introduction
  
  1:48
- 2.
  
  1.1 MCP Overview - What is Model Context Protocol?
  
  4:36
- 3.
  
  1.2. Get Code on Github
  
  1:56
- 4.
  
  SECTION 2 START - Build Your Own MCP Server - 2.1 Introduction to MCP Server and 1 Minute Preview of
  
  1:17
- 5.
  
  2.2 Install Claude for Desktop
  
  1:11
- 6.
  
  2.3 Install Python, UV and VS Code
  
  2:52
- 7.
  
  2.4 Setup Project Directories and Files
  
  1:48
- 8.
  
  2.5 MCP Server Python Code Walkthrough
  
  1:35
- 9.
  
  2.6 Connect Server to Claude Desktop and Test Your Server
  
  12:59
- 10.
  
  SECTION 3 START - Build Your Own MCP Client (Using Python + Google Gemini API) - 3.1 Quick Recap and
  
  2:24
- 11.
  
  3.2 How to Get Free Gemini API Key
  
  1:51
- 12.
  
  3.3 Setup Project Directories, Files & Install Google-GenAI SDK
  
  2:14
- 13.
  
  3.4 MCP Client Python Code Walkthrough
  
  10:33
- 14.
  
  3.5 Test Your MCP Client with Your MCP Server
  
  2:26
- 15.
  
  SECTION 4 - Containerize your MCP Server Using Docker
  
  14:41
- 16.
  
  SECTION 5 - Simplify client code with LangGraph & LangChain
  
  11:28
- 17.
  
  SECTION 6 START Build MCP Client with Multiple Server Support: 6.1 Introduction
  
  0:36
- 18.
  
  6.2 Let's look at the config.json file
  
  0:36
- 19.
  
  6.3 Demo - MCP Client with Multiple MCP Servers
  
  4:10
- 20.
  
  6.4 MCP Client (with json config) Code Walk Through - Part 1
  
  2:23
- 21.
  
  6.5 Why Choose Gemini 2.0 Flash and not the Pro Models
  
  0:47
- 22.
  
  6.6 MCP Client (with json config) Code Walk Through - Part 2 (continued)
  
  3:52
- 23.
  
  6.7 How to use existing MCP servers from MCP Github (example uses "fetch" server)
  
  3:57
- 24.
  
  SECTION 7 START - Server Sent Events - MCP Server and Clients using SSE - 7.1 Introduction
  
  0:58
- 25.
  
  7.2 Quick Recap, What is STDIO and SSE?
  
  2:12
- 26.
  
  7.3 Setup Directories, Clone GitHub Code (git pull only, if done at course start)
  
  1:35
- 27.
  
  7.4 Setup Virtual Environment and Dependencies
  
  2:27
- 28.
  
  7.5 MCP SSE Server Code Walkthrough
  
  8:57
- 29.
  
  7.6 MCP SSE Client Code Walkthrough
  
  14:10
- 30.
  
  7.6 Dockerfile Code (for MCP Server) Walkthrough
  
  0:55
- 31.
  
  7.7 Test your MCP SSE Server and Client Locally
  
  2:36
- 32.
  
  SECTION 8 START - Deploying MCP Server to Google Cloud Platform - 8.1 Create a new Gmail Account (if
  
  2:34
- 33.
  
  8.2 Create a Google Cloud Project
  
  2:48
- 34.
  
  8.3 Install and Setup Google Cloud Command Line Interface (gcloud CLI)
  
  5:33
- 35.
  
  8.4 Build Docker Image for Google Cloud
  
  1:02
- 36.
  
  8.5 Deploy MCP SSE Server to Google Cloud Run
  
  0:58
- 37.
  
  8.6 Test MCP SSE Server on Google Cloud
  
  2:04
- 38.
  
  SECTION 9 START - STREAMABLE HTTP MCP SERVER - 9.1 Quick MCP Recap and What is Streamable HTTP
  
  3:52
- 39.
  
  9.2 Overview of Streamable HTTP using Sequence Diagram
  
  2:54
- 40.
  
  9.3 Initialisation Phase
  
  2:12
- 41.
  
  9.4 Client Requests Phase of Streamable HTTP
  
  1:46
- 42.
  
  9.5 MCP Client Notifications and Responses
  
  0:45
- 43.
  
  9.6 Client Listening to Messages from the Server in Streamable HTTP
  
  1:19
- 44.
  
  9.7 Session handling in Streamable HTTP
  
  1:57
- 45.
  
  9.8 External Resource for keeping up to date
  
  1:57
- 46.
  
  SECTION 9.2 START Streamable HTTP MCP Server - Step-by-step demo
  
  12:02
- 47.
  
  SECTION 9.3 START Streamable HTTP MCP Client using Gemini and Google ADK - Introduction
  
  6:17
- 48.
  
  9.3.2 Code overview
  
  2:49
- 49.
  
  9.3.3 MCP Client User Interface - cmd.py
  
  4:30
- 50.
  
  9.3.4 MCP Client Implementation - client.py
  
  5:11
- 51.
  
  9.3.5 MCP Agent Implementation - agent.py
  
  8:33
- 52.
  
  9.3.6 Code - MCP config json file
  
  2:04
- 53.
  
  9.3.7 Code for Utilities & STDIO Server + Environment Setup
  
  8:17
- 54.
  
  SECTION 10 START - Streamlit User Interface for MCP Client - 10.1 Streamlit UI MCP Client Overview
  
  2:35
- 55.
  
  10.2 Streamlit UI Demo
  
  2:39
- 56.
  
  10.3 Comparing our UI with Claude Desktop!
  
  0:26
- 57.
  
  10.4 Setup Directories (skip if done earlier)
  
  0:54
- 58.
  
  10.5 Setup Google Gemini API Key (verify again if done earlier)
  
  1:37
- 59.
  
  10.6 Create Virtual Environment and Install Dependencies
  
  1:39
- 60.
  
  10.7 Get Streamlit UI Code
  
  1:00
- 61.
  
  10.8 Streamlit App Imports and State Initialisation
  
  3:35
- 62.
  
  10.9 Code for Utility Functions
  
  1:15
- 63.
  
  10.10 Streamlit App Sidebar Code
  
  3:21
- 64.
  
  10.11 Building the Main Chat UI for the App
  
  3:04
- 65.
  
  10.12 Core Logic For Query Handling
  
  1:35
- 66.
  
  10.13 Trigger Logic for Send and Other Chat Buttons
  
  1:41
- 67.
  
  10.14 Streamlit App MCP Client Code
  
  7:04
- 68.
  
  10.15 theailanguage_config.json and other files
  
  1:35
- 69.
  
  10.16 Running the Streamlit UI
  
  0:18
- 70.
  
  SECTION 11 START - A2A or Agent to Agent Protocol - Lifecycle - 11.1 Introduction
  
  2:56
- 71.
  
  11.2 Why A2A Protocol - A2A vs MCP
  
  2:09
- 72.
  
  11.3 Discovery - A2A Client, A2A Server, Agent Card
  
  2:43
- 73.
  
  11.4 Initiation - Tasks, Messages, Parts
  
  1:53
- 74.
  
  11.5 Processing - Artifacts, Streaming, Push Notifications, Non-streaming
  
  1:42
- 75.
  
  11.6 Interaction - input-required state
  
  0:22
- 76.
  
  11.7 Completion Flow & Summary of A2A
  
  1:35
- 77.
  
  SECTION 12 START - Building your own A2A Client and A2A Server
  
  18:40
- 78.
  
  SECTION 13 START - Build your own A2A Agent with Google Agent Development Kit (ADK) - Introduction
  
  3:56
- 79.
  
  13.2 Quick Demo of the A2A Agent
  
  1:43
- 80.
  
  13.3 Code and Gemini API Key Setup
  
  2:57
- 81.
  
  13.4 Run your A2A Agent Server and Client
  
  1:53
- 82.
  
  13.5 Agent Code Walkthrough
  
  4:41
- 83.
  
  13.6 Agent Taskmanager Code Walkthrough
  
  3:36
- 84.
  
  13.7 Agent main.py code walkthrough
  
  2:37
- 85.
  
  13.8 Client code walkthrough - part 1
  
  1:32
- 86.
  
  13.9 Client code walkthrough - part 2 - Manual Agent Discovery
  
  0:50
- 87.
  
  13.10 Client code walkthrough - part 3
  
  2:06
- 88.
  
  13.11 App code walkthrough
  
  3:25
- 89.
  
  13.12 Quick Recap of Code Architecture
  
  0:48
- 90.
  
  13.13 Models code walkthrough
  
  2:35
- 91.
  
  SECTION 14 START - A2A with Multiple Agents
  
  8:28
- 92.
  
  SECTION 15 START - Connect 3 Agents with A2A
  
  4:28
- 93.
  
  15.2 A2A Architecture Recap
  
  1:40
- 94.
  
  15.3 Agent Discovery and Registry
  
  4:19
- 95.
  
  15.4 Orchestrator Host Agent
  
  12:36
- 96.
  
  15.5 Orchestrator Taskmanager
  
  1:38
- 97.
  
  15.6 Orchestrator entry python script
  
  3:10
- 98.
  
  15.7 Greeting agent
  
  8:00
- 99.
  
  15.8 Summary and other files, folders
  
  1:03
- 100.
  
  15.9 Setup the code & API Key
  
  3:42
- 101.
  
  15.10 Running the agents and client
  
  2:39

Nível iniciante

Nível intermediário

Nível avançado

Todos os níveis

Estudantes

Sobre este curso

Class Overview: In this hands-on coding course, you'll learn about Model Context Protocol (MCP) and Agent-to-Agent (A2A) Protocol-the foundational technologies that enable AI agents to communicate, collaborate, and access external tools. You'll build 5 MCP Clients and 3 MCP Servers from scratch, create a Streamlit-based UI for your client in Python, and deploy your MCP Server to Google Cloud using Server Sent Events (SSE). To top it off, we'll use a free Gemini API key from Google, so you can explore these powerful technologies without any cost for AI models.

You'll also implement Agent-to-Agent (A2A) Protocol to connect multiple agents, and build a Host Orchestrator Agent that integrates both MCP and A2A using Google’s Agent Development Kit (ADK).

What You Will Learn

Agent-to-Agent (A2A) Protocol: Connect 3 agents and build a host orchestrator using A2A + MCP.
Build 5 MCP Clients & 3 MCP Servers from Scratch – full working code included.
Design a UI in Streamlit for your MCP Client in Python.
Deploy your MCP Server on Google Cloud using Server Sent Events (SSE).
Use Google’s Gemini API (free key) to run AI-powered agents at no cost.
Work with Python, Gemini, LangGraph, SSE, Streamlit on macOS.

Why You Should Take This Class

MCP allows AI agents to interface with external tools and APIs.
A2A enables agent collaboration across technical and organizational boundaries.
These protocols form the communication backbone for modern autonomous agent systems.
I’ve been building and teaching A2A and MCP-based agent systems that generate code and perform complex tasks - and I’m excited to guide you through building them too.

Who This Class is For

Developers looking to build MCP clients, servers, and user interfaces.
AI agent developers, LangChain developers, and software engineers.
Entrepreneurs and product builders seeking a QuickStart on MCP + A2A.

Materials / Resources

This course is recorded on macOS - access to a macOS machine is highly recommended.
Windows instructions will be provided where relevant (ETA: end of July, as part of Github Repositories).
Linux instructions are currently not available and not under plan to be added later.
You’ll get access to a GitHub repo with all project code
MCP Server Code - https://github.com/theailanguage/terminal_server
MCP Client Code - https://github.com/theailanguage/mcp_client
A2A Code - https://github.com/theailanguage/a2a_samples
MCP and A2A are evolving protocols - breaking changes may occur. The instructor will then try to update the codes to reflect the changes as soon as possible
Basic Python knowledge is required.
Familiarity with LLMs (like Gemini or Claude) is required.

Conheça seu professor

Kartik Marwah

Teaching AI, Agents, MCP, A2A

Professor

Visualizar o perfil completo

Habilidades relacionadas

IA para desenvolvimento Desenvolvimento Linguagens de programação Python

Level: Advanced

Projeto prático de curso

Throughout the course, you will build a series of real-world projects designed to help you master the Agent-to-Agent (A2A) Protocol, Model Context Protocol (MCP), and related tools. Each project comes with full working code and step-by-step guidance.

Project List

Build 5 MCP Clients from Scratch
Create five MCP clients that connect to an MCP Server, send and receive structured messages.
Build 3 MCP Servers from Scratch
Implement three custom MCP Servers with different capabilities, enabling tool execution.
Build a Streamlit UI for MCP Client in Python
Design a user-friendly web interface for your MCP Client using Streamlit, allowing interaction with the protocol using a UI.
Deploy MCP Server to Google Cloud using SSE
Learn to deploy your MCP Server using Server Sent Events (SSE) on Google Cloud, enabling remote MCP Servers.
Use Gemini API with Free Key
Integrate Google’s Gemini API (using a free key) to power your agents without needing to pay for model access while learning.
Connect 3 Agents using A2A Protocol
Implement the Agent-to-Agent (A2A) Protocol to connect three independent agents and enable message-based coordination.
Build a Host Orchestrator Agent using A2A + MCP
Create an advanced orchestrator agent that coordinates tasks across other agents using both A2A and MCP protocols.

Nota do curso

Por que fazer parte da Skillshare?

Faça cursos premiados Skillshare Original

Cada curso possui aulas curtas e projetos práticos

Sua assinatura apoia os professores da Skillshare

Aprenda em qualquer lugar

Faça cursos em qualquer lugar com o aplicativo da Skillshare. Assista no avião, no metrô ou em qualquer lugar que funcione melhor para você, por streaming ou download.

MCP & A2A - Model Context Protocol & Agent to Agent Protocol

Kartik Marwah, Teaching AI, Agents, MCP, A2A

Assista a este curso e milhares de outros

Assista a este curso e milhares de outros

Aulas neste curso

1.

Course Introduction

1:48

2.

1.1 MCP Overview - What is Model Context Protocol?

4:36

3.

1.2. Get Code on Github

1:56

4.

SECTION 2 START - Build Your Own MCP Server - 2.1 Introduction to MCP Server and 1 Minute Preview of

1:17

5.

2.2 Install Claude for Desktop

1:11

6.

2.3 Install Python, UV and VS Code

2:52

7.

2.4 Setup Project Directories and Files

1:48

8.

2.5 MCP Server Python Code Walkthrough

1:35

9.

2.6 Connect Server to Claude Desktop and Test Your Server

12:59

10.

SECTION 3 START - Build Your Own MCP Client (Using Python + Google Gemini API) - 3.1 Quick Recap and

2:24

11.

3.2 How to Get Free Gemini API Key

1:51

12.

3.3 Setup Project Directories, Files & Install Google-GenAI SDK

2:14

13.

3.4 MCP Client Python Code Walkthrough

10:33

14.

3.5 Test Your MCP Client with Your MCP Server

2:26

15.

SECTION 4 - Containerize your MCP Server Using Docker

14:41

16.

SECTION 5 - Simplify client code with LangGraph & LangChain

11:28

17.

SECTION 6 START Build MCP Client with Multiple Server Support: 6.1 Introduction

0:36

18.

6.2 Let's look at the config.json file

0:36

19.

6.3 Demo - MCP Client with Multiple MCP Servers

4:10

20.

6.4 MCP Client (with json config) Code Walk Through - Part 1

2:23

21.

6.5 Why Choose Gemini 2.0 Flash and not the Pro Models

0:47

22.

6.6 MCP Client (with json config) Code Walk Through - Part 2 (continued)

3:52

23.

6.7 How to use existing MCP servers from MCP Github (example uses "fetch" server)

3:57

24.

SECTION 7 START - Server Sent Events - MCP Server and Clients using SSE - 7.1 Introduction

0:58

25.

7.2 Quick Recap, What is STDIO and SSE?

2:12