Teach MCP - Learning to make MCP servers for data analysis tasks
The purpose of the repo is to teach you how to make MCP servers. The target audience are non-technical with no real coding experience.
You must be able to:
- Download and install:
- Python and the dependencies in requirements.txt.
- An MCP host application like VS Code or Claude Desktop. Licenses/subscriptions may apply.
- Git, to be able to clone and use this repo.
- You will need a GitHub account.
- You should set up Git for use with VS Code.
- Be able to read and understand the basic technical concepts as I explain them. You might need to consult with other resources.
- GitHub tutorial for beginners.
- Anthropic MCP docs.
- VS Code docs.
- Python docs.
- Be patient with me. This is the first time I'm teaching something based in Python. I'm putting this out there for free.
License
This repo uses a custom non-commercial license.
You are permitted to use the contents of this repo and derivatives thereof for non-commercial, personal use only. Please enjoy the materials, but also please respect my intellectual property rights.
AI References
- The code and samples in this repo was created with assistance from Claude Opus 4.0 and Claude Sonnet 4.0.
- The docs in this repo were created with assistance from Gemini 2.5 Pro.
- No AI was used in the creation of creative materials or content.
Disclaimer
I do not consider myself an "AI Expert". I am trying to just share what I learned because I find it useful and interesting. The code, information, and examples in this repo are provided as-is without any warranties or guarantees. You are responsible for your own due-diligence and the code you execute or the information that you consume. Nothing I provide should be taken as advice in an official capacity.
Be critical. Be skeptical. Be humble.