Loading video player...
MCP Client Masterclass with full code and sample MCP server to build an MCP Client from scratch in two ways (Updated for 2026) - Udemy Course (get completion certificate, practice questions, Q&A) https://theailanguage.com/best_price?course=mcp_a2a --------------------------------------------------------------------------------------------- 1. MCP Library’s client class (without agent, raw MCP Client) 2. Using Google Agent Development Kit (ADK) Framework’s McpToolset Class (with Agent, to query agent’s to use MCP Servers) 👉 Code used in this video (02_mcp_adk_client): https://theailanguage.com/onlySubscribers?id=mcp-v2-server-examples_vid_ZsqDLpv-GXE_No_list_PL6tW9BrhiPTCDteflzehKS6Cn3a79-iCs&site=github Please subscribe, allow pop-ups, and log in to access our GitHub repos. 🎓 UDEMY COURSES WITH EXCLUSIVE DISCOUNTS Sign in at 👉 https://theailanguage.com to access special discounted links with coupon codes. 📘 Course 1: MCP & A2A Udemy Course ✔ Completion certificate ✔ Practice questions ✔ Q&A support 📗 Course 2: Google ADK Udemy Course ✔ Completion certificate ✔ Q&A support ✔ Bite-sized, easy-to-follow lectures In this complete hands-on tutorial, you will learn how to build a production-style MCP Client using the latest stable Model Context Protocol (v1), integrate it with Google ADK, and connect to multiple STDIO-based MCP servers. We cover: • Raw MCP client implementation using JSON-RPC over STDIO • Lifecycle management with AsyncExitStack • Server discovery via config.json • Tool listing and dynamic tool invocation • Clean session shutdown • Building a sample Echo MCP Server • Integrating MCP with Google Agent Development Kit (ADK) • Using McpToolset inside an LLM Agent • Managing sessions with InMemoryRunner • Running an interactive CLI chat loop You will clearly understand the difference between: • A low-level MCP Client Manager (custom implementation) • ADK’s MCP Toolset abstraction inside an agent framework This is ideal for developers building: • Custom MCP clients • AI agent systems with external tool execution • Multi-server MCP integrations • CLI-based AI tools • Production MCP workflows Full working code included (Client + Server + ADK integration). ⸻ 📚 Chapters 00:00 Introduction 01:27 Quick demo of what we’ll build 04:27 Getting the code 04:57 Code Setup 07:22 MCP Client Code - manager.py 11:10 MCP Client Code - Server Discovery from config 12:45 MCP Client Code - Server Connection 15:58 MCP Client Code - Listing all tools 17:10 MCP Client Code - Calling a tool 18:52 MCP Client Code - Session Cleanup 19:56 MCP Server - simple echo server used in this lecture 21:40 ADK + MCP Integration - Intro and Step 1 - Loading the Server Config 24:47 ADK Integration Step 2 - Connect to MCP Servers 27:54 ADK Integration Step 3 - Initialise the Agent With McpToolset 28:45 ADK Integration Step 4 - Initialise the runner 29:28 ADK Integration Step 5 - Start a conversation Session 30:51 ADK Integration Step 6 - Start an Interactive Chat Loop 35:45 MCP Client Manager vs ADK command line integration ⸻ 🔧 What You’ll Master • JSON-RPC communication over STDIO • Multiplexing multiple MCP servers • Clean async resource management • Agent-driven tool execution • ADK session management • Tool discovery patterns • CLI AI application architecture This tutorial bridges low-level MCP internals and high-level agent frameworks — giving you full clarity and architectural understanding. ⸻ 🧠 Tech Stack • Model Context Protocol (MCP) • Google Agent Development Kit (ADK) • Gemini Models • Async Python • STDIO Transport • JSON-RPC ⸻ 🔎 Who This Is For • AI infrastructure developers • Agent framework builders • Backend engineers building tool-enabled LLM systems • Developers moving from MCP v1 to v2 • Anyone serious about mastering MCP clients ⸻ #️⃣ Hashtags #MCPV2#ModelContextProtocol #MCPClient #GoogleADK #AgentDevelopmentKit #LLMAgent #GeminiAPI #AIAgents #JSONRPC #PythonAsync #AIInfrastructure #ToolCalling #STDIO #MCPServer #AIDevelopment #AIAutomation ⸻