This is part 3 of the Build Series. Part 1 — Part 2
There was a moment last week when I sat back and looked at the Vault and thought: this is too much work to use.
Not broken. Not ugly. But very inconvenient in a way that compounds.
A subscriber wants to use one of my blueprints. So they stop what they are doing, open a browser, log in, run a search, read the result, copy the prompt, go back to Claude, and paste it in. Seven steps. Every single time.
I had built something useful that was also annoying to reach. And in product, annoying to reach eventually means not used at all.
So I stopped adding features and asked a simpler question: what if they never had to leave Claude?
That is what I built last week. 🚀
The Real Friction Problem
When I launched the Vault, I thought the hard part was the content. Getting 62 blueprints cleaned, structured, and searchable. And it was hard. Parts 1 and 2 of this series are basically a documentation of that chaos.
But the content was never the problem. The problem was that my subscribers had to come to it. Open a tab, log in, search, copy, paste, go back to work.
That is a lot of steps for something that is supposed to speed you up.
What I actually needed was for the logic to go to them. Right inside the tool they are already using. No tab switching. No copy pasting. Just: ask Claude, get the blueprint, keep building.
That is what MCP makes possible. And that is what I built.