Build Your First MCP Server From Scratch - Model Context Protocol [Module 2]

Ask Claude to fix one typo in your Google Drive draft and you can end up with two files that share a name. This video builds an MCP server in Python that edits the real file. 🎓 Want to learn Claude Code with me, live? Join the Claude Code Builder cohort on Maven → https://maven.com/genaiunplugged/claude-code-builder This covers lessons 11 to 15 of the MCP Masterclass, module 2. You build a small server called Scribe, test it with a proof script, and then add resources and a prompt. 🎯 What You Will Learn - Why the official Google Drive server can find your draft but has no tool that reads it, and why Claude's built-in Drive connector can only save a copy - How to write a plain Python function and make it an MCP tool with one line above it - How the tool name and description come from your function name and your docstring - How to keep your folder id and key in the environment, so you can share the file with a friend and give away nothing - How to run 6 tests with a prove it script, and why the one failing test taught the most - How to add 2 resources and 1 prompt, so Claude can see your folder and reuse a saved editing job 🧰 What You Need - Python 3 on your computer - Claude Desktop - A Google Drive folder that you share with the server - The module 2 checkpoint, which holds 3 sample drafts and a pinned package list - Windows users, your command lines are in the checkpoint notes 🧪 The 6 Tests You Run - Go and look: Drive itself names your server as the one that changed the file - Run the proof script, as often as you like - Missing words change nothing - A missing draft returns a plain sentence, so Claude can explain the problem - Scribe cannot make a file, and it says so - A second app uses your server with no new code 🗺️ Where You End Up Your server finishes with 2 tools, 2 resources and 1 prompt. Fixing a draft costs you one sentence, and you never open the document. In module 3 we make Scribe safe to trust: it stays inside its folder, it asks before it writes over anything, and it remembers what you told it. 📦 Module 2 checkpoint (working code, pinned packages, setup check): https://www.genaiunplugged.com/academy/mcp-masterclass/checkpoints/scribe-checkpoint-m2.zip 📖 Every lesson page, code and the other checkpoints: https://www.genaiunplugged.com/courses/mcp/ 📺 CHAPTERS: 0:00 Lesson 11: Why the official Google Drive server cannot fix one typo 5:43 Lesson 12: Write your first MCP server: one tool that edits a Drive draft 13:13 Lesson 13: 6 tests for your MCP server, and the one that fails 19:05 Lesson 14: Stop retyping your editing job: resources and prompts 25:14 Lesson 15: The hole in the map is gone: module 2 in five points 🔗 RESOURCES Model Context Protocol official site and docs → https://modelcontextprotocol.io Official MCP Python SDK on GitHub → https://github.com/modelcontextprotocol/python-sdk Claude Desktop download → https://claude.ai/download Python downloads → https://www.python.org/downloads/ 📥 Helpful Links: 🛠️ Check my AI Systems store → https://store.genaiunplugged.com/ ▶️ Watch the Full n8n Zero to Hero Course Playlist → https://tinyurl.com/3rw3x6hy 🌐 Subscribe to GenAI Unplugged Substack → https://genaiunplugged.substack.com/ 🎓 Learn with me on Maven → https://maven.com/genaiunplugged 🤝 Let's connect on LinkedIn → https://www.linkedin.com/in/dheerajsharma14/ 📚 Browse all my courses → https://genaiunplugged.com/ 🙌 A like is what tells YouTube to show this to the next person looking for the same fix. Subscribe if you want a new AI build every week. Tell me in the comments which of the 6 tests failed on your run and what your Scribe error sentence said, because I read every comment. ⚠️ DISCLAIMER This video is educational. Every setup I show uses the provider's own documented features. Nothing here bypasses billing, cracks licensing, or works around anyone's terms of service. I am not affiliated with any of the tools or providers shown, and there are no affiliate links in this description. Free tiers, model availability and rate limits change often, so check the official docs before you follow along. Your API keys are yours to protect. Never commit one to a repo or leave one on screen. Do not send confidential or client data through a third party model provider. Follow at your own risk. I accept no liability for any cost or issue that comes out of this tutorial. #MCP #ModelContextProtocol #ClaudeAI