Follow dis steps to begin to dey use these resources:
- Fork di Repository: Click
- Clone di Repository:
git clone https://github.com/microsoft/mcp-for-beginners.git - Join Di
Arabic | Bengali | Bulgarian | Burmese (Myanmar) | Chinese (Simplified) | Chinese (Traditional, Hong Kong) | Chinese (Traditional, Macau) | Chinese (Traditional, Taiwan) | Croatian | Czech | Danish | Dutch | Estonian | Finnish | French | German | Greek | Hebrew | Hindi | Hungarian | Indonesian | Italian | Japanese | Kannada | Khmer | Korean | Lithuanian | Malay | Malayalam | Marathi | Nepali | Nigerian Pidgin | Norwegian | Persian (Farsi) | Polish | Portuguese (Brazil) | Portuguese (Portugal) | Punjabi (Gurmukhi) | Romanian | Russian | Serbian (Cyrillic) | Slovak | Slovenian | Spanish | Swahili | Swedish | Tagalog (Filipino) | Tamil | Telugu | Thai | Turkish | Ukrainian | Urdu | Vietnamese
You prefer to Clone for your machine?
Dis repository get pass 50 languagetranslation wey dey inside am wey go make di download size big well well. If you want clone without di translations, use sparse checkout:
Bash / macOS / Linux:
git clone --filter=blob:none --sparse https://github.com/microsoft/mcp-for-beginners.git cd mcp-for-beginners git sparse-checkout set --no-cone '/*' '!translations' '!translated_images'CMD (Windows):
git clone --filter=blob:none --sparse https://github.com/microsoft/mcp-for-beginners.git cd mcp-for-beginners git sparse-checkout set --no-cone "/*" "!translations" "!translated_images"Dis one go give you everything wey you need to complete di course with beta download speed.
Welcome to your waka enter di Model Context Protocol! If you don ever wonder how AI applications dey take dey talk to different tools and services, you dey yan just about find di fine solution wey dey change how developers dey build intelligent systems.
Think say MCP na universal translator for AI applications - just like how USB ports dey make you fit connect any device to your computer, MCP dey make AI models fit connect to any tool or service for one standard way. Whether you dey build your first chatbot or na complex AI workflows you dey do, to sabi MCP go give you power to make beta and flexible applications.
Dis curriculum design am with patience and care for your learning journey. We go start wit simple concepts wey you sabi before and slowly we go build your skills through hands-on practice for your favorite programming language. Every step include clear explanation, practical examples, and plenty encouragement along the way.
By di time we done with this journey, you go get confidence to build your own MCP servers, join dem with popular AI platforms, and sabi how dis technology dey reshaping di future of AI development. Make we start dis exciting journey together!
Dis curriculum align wit MCP Specification 2025-11-25 (dis na di latest stable release). Di MCP specification dey use date-based versioning (YYYY-MM-DD format) so dat everybody go fit track protocol version clear.
These resources go beta as your understanding grow, but no pressure say you must read everything sharp sharp. Start with di ones wey you dey interested for!
- 📘 MCP Documentation – Na your go-to resource for step-by-step tutorials and user guides. Di documentation dey write am for beginners, so e get clear examples wey you fit follow at your own pace.
- 📜 MCP Specification – Think am as your full reference manual. As you dey waka through di curriculum, you go dey come back here to check details and explore advanced features.
- 📜 MCP Specification Versioning – E get info about protocol version history and how MCP dey use date-based versioning (YYYY-MM-DD format).
- 🧑💻 MCP GitHub Repository – You go find SDKs, tools, and code samples for many programming languages here. E be like treasure house of practical examples and ready-to-use components.
- 🌐 MCP Community – Join other learners and experienced developers to discuss about MCP. E good community wey questions dey welcome and knowledge dey shared freely.
By di end of dis curriculum, you go get confidence and e go excite you for your new skills. Here na wetin you go fit do:
• Understand MCP fundamentals: You go sabi wetin be Model Context Protocol and why e dey change how AI applications dey waka together, using simple examples and analogies wey make sense.
• Build your first MCP server: You go fit create working MCP server for your preferred programming language, start with easy examples and build your skill step by step.
• Connect AI models to real tools: You go learn how to link AI models to real services, make your applications get beta new power.
• Implement security best practices: You go sabi how to keep your MCP implementations safe, make your applications and your users dey protected.
• Deploy with confidence: You go know how to carry your MCP projects from development enter production, with practical deployment tactics wey work for real world.
• Join the MCP community: You go become part of growing community of developers wey dey shape di future of AI application development.
Before we dive inside MCP details, make sure you dey comfortable with some start-up concepts. No worry if you no be expert for these areas - we go explain everything wey you need to sabi as we dey go!
Think say protocol na like rules for how person go talk. If you call your friend, both una sabi say make una talk "hello" wen una pick, take turns to talk, and talk "goodbye" wen una finish. Computer programs also need rule like dis to fit talk well.
MCP na protocol - na set of rules wey dey help AI models and applications get good "conversations" with tools and services. Just like how conversation rules dey make human talk beta, MCP dey make AI application communication more reliable and strong.
You dey already use client-server relationship everyday! When you use web browser (na di client) to visit website, you dey connect to web server wey go send you di page. Di browser sabi how to ask for info, and di server sabi how to answer.
For MCP, we get similar relation: AI models dey act as clients wey dey ask for info or do something, while MCP servers go provide these tings. E be like assistant (di server) wey AI fit ask to do certain jobs.
Imagine if every car maker get different shape for fuel fill point - you go need different adapter for each car! Standardization na to agree on common way so tings go fit work together smoothly.
MCP na di standardization for AI applications. Instead make every AI model need different code for every tool, MCP create one universal way for dem to dey talk. This one mean say developers fit build tools once and e go work with many different AI systems.
Your MCP journey get plan well well to build your confidence and skills step by step. Each phase go show new concepts and also remind you wetin you don learn before.
Na here your journey dey start! We go introduce you to MCP concepts using examples wey you sabi well well. You go sabi wetin MCP be, why e dey, and how e take fit inside bigger AI development world.
• Module 0 - Introduction to MCP: We go start to explore wetin MCP be and why e mean plenty for modern AI applications. You go see real examples of MCP work and understand how e solve common problems wey developers dey face.
• Module 1 - Core Concepts Explained: Here you go learn important parts of MCP. We go use many analogies and visual examples to make sure you understand these concepts well.
• Module 2 - Security in MCP: Security fit look like hard matter, but we go show you how MCP get safety features inside and teach you best practices wey go protect your applications from day one.
Now di real beta enjoyment don start! You go get hands-on experience to build real MCP servers and clients. No wahala - we go start easy and guide you for every step.
Dis module get plenty hands-on guides wey go make you practice for your favorite programming language. You go create your first server, build client wey go connect to am, plus join popular development tools like VS Code.
Every guide get full code examples, troubleshooting tips, and explanation why we dey do specific design choices. By the time you finish dis phase, you go get working MCP implementations wey you fit dey proud of!
With basics we don sabi well, you ready to explore more better MCP features. We go cover practical ways to implement am, debugging methods, and advanced topics like multi-modal AI integration.
You go still learn how to scale your MCP implementations for production use plus join cloud platforms like Azure. Dis modules go prepare you to build MCP solutions wey fit handle real-world wahala.
Final phase na to join MCP community plus specialize for areas wey you like pass. You go learn how to contribute to open-source MCP projects, use advanced authentication patterns, and build complete database-integrated solutions.
Module 11 get special talk - na full 13-lab hands-on learning path wey dey teach you how to build production-ready MCP servers with PostgreSQL integration. E be like capstone project wey join everything wey you don learn!
| Module | Topic | Description | Link |
|---|---|---|---|
| Module 0-3: Fundamentals | |||
| 00 | Introduction to MCP | Overview of the Model Context Protocol and its significance in AI pipelines | Read more |
| 01 | Core Concepts Explained | Deep talk on core MCP concepts | Read more |
| 02 | Security in MCP | Security threats and best ways to protect | Read more |
| 03 | Getting Started with MCP | Environment setup, basic servers/clients, integration | Read more |
| Module 3: Building Your First Server & Client | |||
| 3.1 | First Server | Build your first MCP server | Guide |
| 3.2 | First Client | Develop simple MCP client | Guide |
| 3.3 | Client with LLM | Join large language models | Guide |
| 3.4 | VS Code Integration | Use MCP servers for VS Code | Guide |
| 3.5 | stdio Server | Build servers using stdio transport | Guide |
| 3.6 | HTTP Streaming | Do HTTP streaming for MCP | Guide |
| 3.7 | AI Toolkit | Use AI Toolkit with MCP | Guide |
| 3.8 | Testing | Test your MCP server implementation | Guide |
| 3.9 | Deployment | Deploy MCP servers to production | Guide |
| 3.10 | Advanced server usage | Use advanced servers for better feature use and improved system design | Guide |
| 3.11 | Simple auth | Chapter wey show you auth from the start and RBAC | Guide |
| 3.12 | MCP Hosts | Configure Claude Desktop, Cursor, Cline, and other MCP hosts | Guide |
| 3.13 | MCP Inspector | Debug and test MCP servers with Inspector tool | Guide |
| 3.14 | Sampling | Use sampling to work with client | Guide |
| 3.15 | MCP Apps | Build MCP Apps | Guide |
| Module 4-5: Practical & Advanced | |||
| 04 | Practical Implementation | SDKs, debugging, testing, reusable prompt templates | Read more |
| 4.1 | Pagination | Manage big result sets with cursor-based pagination | Guide |
| 05 | Advanced Topics in MCP | Multi-modal AI, scaling, enterprise use | Read more |
| 5.1 | Azure Integration | MCP join with Azure | Guide |
| 5.2 | Multi-modality | Work with many modalities | Guide |
| 5.3 | OAuth2 Demo | Do OAuth2 authentication | Guide |
| 5.4 | Root Contexts | Understand and do root contexts | Guide |
| 5.5 | Routing | MCP routing ways | Guide |
| 5.6 | Sampling | Sampling methods for MCP | Guide |
| 5.7 | Scaling | Scale MCP implementations | Guide |
| 5.8 | Security | Advanced security things | Guide |
| 5.9 | Web Search | Implement web search ability | Guide |
| 5.10 | Realtime Streaming | Build realtime streaming features | Guide |
| 5.11 | Realtime Search | Do realtime search | Guide |
| 5.12 | Entra ID Auth | Authentication with Microsoft Entra ID | Guide |
| 5.13 | Foundry Integration | Join with Azure AI Foundry | Guide |
| 5.14 | Context Engineering | Ways to do better context engineering | Guide |
| 5.15 | MCP Custom Transport | Custom Transport implementations | Guide |
| 5.16 | Protocol Features | Progress notifications, cancellation, resource templates | Guide |
| 5.17 | Adversarial Multi-Agent Reasoning | Two agents argue opposite sides using shared MCP tools, evaluated by a judge agent | Guide |
| Module 6-10: Community & Best Practices | |||
| 06 | Community Contributions | How to contribute to MCP things | Guide |
| 07 | Insights from Early Adoption | Real-world stories of implementation | Guide |
| 08 | Best Practices for MCP | Performance, fault-tolerance, resilience | Guide |
| 09 | MCP Case Studies | Real implementation examples | Guide |
| 10 | Hands-on Workshop | Build MCP Server with AI Toolkit | Lab |
| Module 11: MCP Server Hands On Lab | |||
| 11 | MCP Server Database Integration | Full 13-lab hands-on learning path for PostgreSQL integration | Labs |
| 11.1 | Introduction | Overview of MCP with database join and retail analytics use case | Lab 00 |
| 11.2 | Core Architecture | Understanding MCP server architecture, database layers, and security patterns | Lab 01 |
| 11.3 | Security & Multi-Tenancy | Row Level Security, authentication, and multi-tenant data access | Lab 02 |
| 11.4 | Environment Setup | Setting up development environment, Docker, Azure resources | Lab 03 |
| 11.5 | Database Design | PostgreSQL setup, retail schema design, and sample data | Lab 04 |
| 11.6 | MCP Server Implementation | Build FastMCP server with database join | Lab 05 |
| 11.7 | Tool Development | Make database query tools and schema introspection | Lab 06 |
| 11.8 | Semantic Search | Build vector embeddings with Azure OpenAI and pgvector | Lab 07 |
| 11.9 | Testing & Debugging | Testing methods, debugging tools, and validation ways | Lab 08 |
| 11.10 | VS Code Integration | Setup VS Code MCP join and AI Chat use | Lab 09 |
| 11.11 | Deployment Strategies | Docker deployment, Azure Container Apps, and scaling issues | Lab 10 |
| 11.12 | Monitoring | Application Insights, logging, performance checking | Lab 11 |
| 11.13 | Best Practices | Make performance better, tighten security, and production tips | Lab 12 |
One of di most exciting part of learning MCP na to see how your code skill dey grow step by step. We don design our code examples to start easy and become more advanced as you understand well. Na so we dey introduce concepts - with code wey easy to understand but show real MCP principles, you go sabi not just wetin dis code dey do, but why e dey structured so and how e take fit for bigger MCP applications.
| Language | Description | Link |
|---|---|---|
| C# | MCP Server Example | View Code |
| Java | MCP Calculator | View Code |
| JavaScript | MCP Demo | View Code |
| Python | MCP Server | View Code |
| TypeScript | MCP Example | View Code |
| Rust | MCP Example | View Code |
| Language | Description | Link |
|---|---|---|
| C# | Advanced Sample | View Code |
| Java with Spring | Container App Example | View Code |
| JavaScript | Advanced Sample | View Code |
| Python | Complex Implementation | View Code |
| TypeScript | Container Sample | View Code |
To get di most out of dis curriculum, you suppose get:
-
Basic knowledge of programming for at least one of these languages: C#, Java, JavaScript, Python, or TypeScript
-
Understanding of client-server model and APIs
-
Familiarity with REST and HTTP concepts
-
(Optional) Background for AI/ML concepts
-
Join our community discussions for support
Dis repository get plenti resources to help you navigate and learn well:
One beta Study Guide dey available to help you navigate dis repository well. Dis visual curriculum map show how all di topics join together and guide you how to use di sample projects well. E dey especially good if you be visual learner wey like to see di big picture.
Di guide get:
- One visual curriculum map wey show all di topics wey dey inside
- Detailed breakdown of each part for di repository
- Guidance on how to use sample projects
- Recommended learning paths for different skill levels
- Extra resources to help your learning journey
We dey maintain one detailed Changelog wey track all important updates to di curriculum materials, so you fit dey up-to-date with di latest improvements and new tins.
- New content wey we add
- Structural changes
- Feature improvements
- Documentation updates
Each lesson for this guide get:
- Clear talk about MCP concepts
- Live code examples for different languages
- Exercises to build real MCP applications
- Extra resources for advanced learners
Make we learn about Model Context Protocol (MCP), one beta new framework wey dem design to standardize interactions between AI models and client applications. For this beginner-friendly session, we go introduce you to MCP and guide you how to create your first MCP server.
JavaScript: https://aka.ms/letslearnmcp-javascript
Congrats! You don already take di first step for one exciting journey wey go expand your programming skills and connect you to the cutting edge of AI development.
By reading dis introduction, you don start to build your MCP knowledge foundation. You sabi wetin MCP be, why e important, and how dis curriculum go support your learning journey. Na big achievement dis and na di start of your skill for dis important technology.
As you dey move through the modules, remember say every expert start as beginner. Di concepts wey fit hard now go become easy as you practice and apply dem. Each small step dey build strong skills wey go serve you throughout your development career.
You dey join one community of learners and experts wey dey passionate about MCP and ready to help others succeed. Whether you get wahala with coding or you dey happy to share your breakthrough, di community dey here to support your journey.
If you jam problem or get questions about building AI apps. Join fellow learners and experienced developers for discussions about MCP. Na supportive community where questions dey welcome and knowledge dey shared freely.
If you get product feedback or find errors while building, visit:
Your MCP journey start now! Begin with Module 0 to enter your first hands-on MCP experiences, or explore di sample projects to see wetin you go dey build. Remember - every expert start from where you dey now, and with patience plus practice, you go shock how far you fit go.
Welcome to di world of Model Context Protocol development. Make we build beta things together!
Dis curriculum dey stronger as learners like you dey contribute! Whether you dey fix small typo, suggest clearer explanation, or add new example, your contributions dey help other beginners succeed.
Thanks to Microsoft Valued Professional Shivam Goyal for contributing code samples.
Di contribution process na welcoming and supportive one. Most contributions need Contributor License Agreement (CLA), but di automated tools go guide you well well.
This complete curriculum dey available under di MIT LICENSE, meaning you fit use, change, and share am freely. Dis support our mission to make MCP knowledge available for developers everywhere.
Dis project dey happy to get contributions and suggestions. Most contributions require you to agree to one Contributor License Agreement (CLA) wey talk say you get di right to, and actually dey grant us the right to use your contribution. For details, check https://cla.opensource.microsoft.com.
When you submit pull request, one CLA bot go automatically decide if you need provide CLA and go decorate di PR properly (for example, status check, comment). Just follow instructions wey the bot go give you. You go only need do am once across all repos wey dey use our CLA.
This project don adopt the Microsoft Open Source Code of Conduct. For more info see di Code of Conduct FAQ or contact opencode@microsoft.com if you get any extra questions or comments.
Ready to start your MCP journey? Begin with Module 00 - Introduction to MCP and take your first steps into di world of Model Context Protocol development!
Our team dey produce other courses! Check am out:
-9333EA?style=for-the-badge&labelColor=E5E7EB&color=9333EA){kind=icon}
-C084FC?style=for-the-badge&labelColor=E5E7EB&color=C084FC){kind=icon}
-E879F9?style=for-the-badge&labelColor=E5E7EB&color=E879F9){kind=icon}
Disclaimer:
Dis document na wetin AI translation service Co-op Translator translate am. Even though we dey try make am correct, abeg sabi say automated translations fit get some errors or mistakes. Di original document wey dey im own language na di correct one to trust. For important tori, e better make human professional translation do am. We nor go hold ourselves responsible for any misunderstanding or wrong meaning wey fit come from dis translation.