Putting It Together: One Capability, Three Doors
The point of learning all three is seeing they are the same capability behind different doors. Build the logic once; open a CLI for people, an API for software, and an MCP server for AI.
Here's the move the whole course was built toward: write a capability once, then put three doors on it.
By the end of this lesson
You'll see how one piece of logic gets exposed through a CLI, an API, and an MCP server at once — and how those three doors stack into a single chain of action.
Take one function — publish_entry(), which validates a draft and saves it. You write it once. Then you wrap it three thin ways, and a person, a program, and a model can all reach the same code without you rebuilding anything underneath.
Same logic, three callers. But they don't only sit side by side — they stack. A flow you'll build often runs straight down: an AI agent calls an MCP tool; that tool's code calls an internal API; behind the API, a deploy step shells out to a CLI command.
Each layer hides the one below it, which is exactly what keeps big systems readable. The model never needs to know there's a CLI at the bottom of the chain, the same way the API never needed to know what it called. That hiding is not an accident of MCP — it's the contract idea from the very first lesson, repeated at every level.
One idea, aimed at the human, the program, and the model in turn.
That's the whole course in a sentence. CLIs, APIs, and MCP were never three subjects to memorize — they're one idea, a clean contract for asking a system to act, pointed at three different callers. Build the capability once, expose whichever doors your callers need, and let them stack when the work calls for it.
Where to go next
From here, Automation & Agents turns the MCP-tool idea into running systems, and Building with Claude shows the agent loop up close. You now have the map; those courses build on it.
Sources — Model Context Protocol, Architecture; earlier lessons cite Anthropic (MCP), Roy Fielding (REST, 2000), the OpenAPI Initiative, and the Bell Labs / Unix shell history.
0
Tap to appreciate
Was this chapter helpful?