Getting Yapship's core loop working happened faster than I expected. Wire a form to Claude, wire Claude to Notion, done — an idea comes in, a draft comes out. If I'm honest, the part I was bracing for as a hard technical problem wasn't hard at all.
The 20% that wasn't
What actually took time was everything around that loop — the stuff a user never consciously notices, right up until the one time it's missing.
- Making sure a half-finished draft doesn't just sit there forever — someone has to notice a writer went quiet for a few days and nudge them back, without pestering them if they're just still writing.
- Making sure the “send this to my publisher” button can only fire once, so nobody accidentally re-sends a live draft and resets a deadline that already went out.
- Making sure the SEO review shows up no matter which button someone hits at the end, so an edit made right before shipping never silently vanishes depending on which choice they picked.
- Making sure an uploaded image link doesn't quietly rot a few hours after publishing, because the storage layer only guarantees that link stays valid for about an hour.
- Deciding, structurally, what actually counts as “an idea” versus “a finished piece of content” — a distinction that felt pedantic on day one and turned out to matter every time we wanted to turn one idea into more than one format.
Even this post
Even this blog didn't skip the pattern. What I thought would be “just add a blog page” turned into a handful of real conversations about what it was actually for, whose data it should touch, and how it should stay separate from everything customers do inside the product. The page itself was quick. Getting those decisions right wasn't.
That's been the whole experience of building Yapship so far. The product is the easy part. It's the little things that decide whether anyone trusts it.