First guides in progress

Learn the Model Context Protocol, one practical guide at a time

MCPRank Guides is a learning hub for MCP: what it is, how to connect servers to the AI tools you already use, how to build your own, and how to do it safely. Plain-language explanations, checked against the official docs.

# Add a remote MCP server to Claude Code
$ claude mcp add --transport http notion https://mcp.notion.com/mcp

# Add a local (stdio) server
$ claude mcp add --transport stdio myserver -- npx -y my-mcp-server

# Check what's connected
$ claude mcp list
Start here

What is MCP, in one paragraph

The Model Context Protocol (MCP) is an open-source standard for connecting AI applications to external systems. The official docs compare it to a USB-C port for AI: one standard way for assistants like Claude or ChatGPT, and developer tools like VS Code and Cursor, to reach files, databases, APIs and workflows. A server exposes tools, resources and prompts. A client inside the AI application connects to it, either as a local process over stdio or remotely over Streamable HTTP.

Learning tracks

Four tracks, from first install to production

Guides are being written and fact-checked now. Each track shows what it will cover and links to its section of the blog.

01

MCP fundamentals

The ideas you need before installing anything.

  • Hosts, clients and servers
  • Tools, resources and prompts
  • Local stdio vs. remote Streamable HTTP
In progressMCP basics →
02

Install a server in your client

Step-by-step setup for the clients people use most.

  • Claude Desktop
  • Claude Code
  • Cursor
  • VS Code
03

Build your own server

From an empty folder to a working server using an official SDK.

  • Choosing an SDK (TypeScript, Python, C#, Go, Rust, Ruby and more)
  • Defining your first tool
  • Testing and debugging locally
04

Security best practices

How to give an AI access to real systems without regretting it.

  • Least-privilege access
  • Handling secrets and tokens
  • Vetting third-party servers
In progressSecurity →
Quick reference

Where each client keeps its MCP config

A cheat sheet while the full guides are being written. Always check the client’s own documentation, because these details can change between versions.

ClientHow to add a serverTop-level key
Claude DesktopSettings → Developer → Edit Config opens claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\). Restart the app after saving.mcpServers
Claude Codeclaude mcp add from the terminal. Project-scoped servers are shared through .mcp.json in the project root. Use /mcp in a session to check status.mcpServers
Cursor.cursor/mcp.json in a project, or ~/.cursor/mcp.json for all projects.mcpServers
VS Code.vscode/mcp.json in your workspace. Tools then become available in the Chat view.servers

Sources: modelcontextprotocol.io, Claude Code docs, Cursor docs, VS Code docs. Checked October 2026.

Security first

Three habits before you connect anything

Grant the minimum

A local server runs with your user permissions. Only give it the folders, repositories or accounts it actually needs.

Keep secrets out of shared files

Project config files are often committed to git. Keep API keys in environment variables or user-level config, not in files your team shares.

Read before you approve

Install servers from sources you trust, and check what each tool call will do before you approve it.

The guides

Browse the guides by topic

Every guide lives on the blog, grouped by topic. New guides appear here as soon as they pass fact-checking.

Go to all guides →

Choosing a server to install? MCPRank is building a directory of MCP servers scored on transparent, published criteria. Rankings are launching soon.

Visit MCPRank →
FAQ

Common questions

Do I need to know how to code to use MCP?

No. Using an existing server usually means adding a short block of JSON to a config file, or running one command in Claude Code. Building your own server does require programming, and the official SDKs make that easier.

What’s the difference between stdio and Streamable HTTP?

They are the two standard MCP transports. With stdio, the client starts the server as a local process and talks to it through standard input and output. With Streamable HTTP, the server runs remotely and the client sends messages to a single HTTP endpoint.

Which languages can I build an MCP server in?

The MCP project lists official SDKs for TypeScript, Python, C#, Go, Rust, Ruby, Java, Swift, PHP and Kotlin, grouped into tiers by feature completeness and maintenance commitment.

Is it safe to install MCP servers?

It can be, if you’re careful. A local server can do anything your user account can do within the access you give it, so only install servers you trust, limit what they can reach, and review actions before approving them.

When will the guides be published?

The first guides are being written and fact-checked now. The fundamentals and client setup tracks will come first.

The first guides are on the way

Until they land, the official MCP documentation is the best place to go deeper.

Open the official MCP docs