WTF is MCP (Model Context Protocol)?

Model Context Protocol is an open standard created by Anthropic that defines how AI assistants connect to external systems.

You can think of it as a universal adapter. Instead of building custom integrations for every API you want to use, developers can build an MCP server that sits between the assistant and the API, standardizing the communication between them.

Core MCP concepts

MCP servers expose three building blocks:

1. Resources: The data supplied by an API or other data source

2. Tools: Functions that can be called by the LLM

3. Prompts: Templates that are written to help users accomplish specific tasks

When these are all put together, they create a clean, consistent way for LLMs to interact with external systems and data.

Examples of MCPs

There are already thousands of MCP servers that exist, and more are created every day. You can even create your own.

Here are a few examples:

• Atlassian, to access Jira and Confluence

• Cloudflare, to get documentation or spin up development environments

• Notion, to connect your workspace

• Stripe, to execute payment processing tasks

• Zapier, to execute automated workflows

When any of these MCPs are installed and enabled, you can interact with these external services directly from your chat client — whether that's ChatGPT, Claude Code, or another MCP-compatible tool. This lets you talk with external systems in plain language, greatly expanding what's possible in your AI workflows.