Claude Code is great at reading and editing your code, but a lot of your work lives outside the repo: issue trackers, databases, docs, and internal APIs. The Model Context Protocol (MCP) is how Claude Code reaches those systems.
Servers give Claude tools
An MCP server is a small program that exposes capabilities over a standard protocol. Once you connect one, Claude Code can call the tools it offers and read the resources it publishes. That means Claude can look up a ticket, query a table, or fetch a document without you copying and pasting the data in.
You connect a server with one command:
claude mcp add --transport http notion https://mcp.notion.com/mcp
After that, the server's tools become available in your session, and Claude decides when to use them based on what you ask.
Tools and resources
A server can offer two kinds of things. Tools are actions Claude can run, like creating an issue or running a query. Resources are pieces of content you can pull in, which you reference with @ mentions.
Why it matters
MCP keeps integrations consistent. Instead of a different plugin format for every service, you add servers the same way every time, and they work across your projects. Servers run locally as a process or connect to a remote endpoint, so you can mix a local database bridge with a hosted SaaS server in the same session.
The rest of this series walks through adding servers, choosing transports and scopes, sharing them with your team, authenticating, and fixing connections when they break.
Comments
Be the first to comment.