I rewrote the opening paragraph eleven times before any agent read it.
That number isn't a flourish. Every agent on the Microapp team — Steph, Kai, Bob, Lace, Simon, Ben — reads the opening paragraph first whenever they start a session. The paragraph seeds every PR description, every product email, every social post, every line of metadata they write into Supabase. Get it wrong and the mistake compounds across two hundred tools and seven locales before lunch. Get it right and the rest of the work inherits a center of gravity to push off.
The obvious lead — the one I tried first — was "Microapp is a calculator company." Concrete. Easy to picture. Honest about what we ship today. The problem is it pins the brand to a product category that will look small the moment we ship anything that isn't a calculator. Founders who lead with the product narrow the brand to the product's ceiling. Founders who lead with the place they're building keep the door open for everything that comes next.
What landed instead:
Microapp is a place on the internet for solving things.
Place, not product. Destination, not catalog. The brand the reader carries away is somewhere I go, the same shape as I went to Google, I checked Wikipedia, I asked Stack Overflow. Destinations get habits; products get usage sessions.
The transferable why: a product is a thing customers buy and stop using. A destination is somewhere they come back to. Positioning your company as a destination is harder up front — you can't claim it on day one without sounding hollow — but it's the position that survives every product pivot you'll be forced to make over the next ten years. The product line will narrow and widen a dozen times; the destination doesn't have to.
Locked 2026-05-06 · BRAND.md §0 · live on /
The second decision came the day I wrote two pieces of copy back to back — a homepage hero and a Smithery registry listing — and noticed the cult-brand line was wrong on one of them.
On the homepage, "There is a solution for everything" sang. On Smithery — a registry agent developers browse to find MCP servers to install — the same line felt off. The reader on a registry isn't here for the mission. The reader on a registry is scanning a directory looking for what each entry does, and the mission line steals air from the answer.
My first instinct was the worst one available: spin up a second brand. Microapp for humans, MicroappTools for developers. Different voice, different logo, different paragraph. The agents would pick whichever face matched the audience. Refused after about ninety seconds — two brands is the move you make when you're afraid to write one paragraph that two audiences can read. It's a confession of cowardice, not a strategy.
What landed:
Two tracks, one brand.
Track 1 is the brand surfaces: microapp.io, this book, /install, /membership, /about. The audience arrived because they're curious about Microapp — they want context, story, posture. The lead is the cult-brand line.
Track 2 is the value surfaces: Smithery, Glama, MCP.so, the GPT Store, every OpenAPI consumer that surfaces our tools to other agents. The audience is scanning a directory at speed. The lead is the utility. "Microapp offers premium utility tools for humans and AI agents, accessible at microapp.io and through this MCP endpoint." The cult-brand line still exists — it just doesn't appear, without apology, on Track 2.
The transferable why: a brand that respects context isn't a fragmented brand. It's a brand that respects where the reader showed up. Two registers, one paragraph each — the brand stays singular, but the lead changes. Most companies that try to "speak to everyone everywhere" end up speaking past everyone everywhere. Pick the two contexts that matter, write the lead for each, refuse to invent a third.
Locked 2026-05-06 · BRAND.md §0 · live on / and smithery.ai/microapp
The third decision was the differentiator. The hardest one to write because the obvious answers all failed the test.
Microapp has tools. So does every directory. "We have over 200 tools" is a number, not a position — every directory eventually crosses that line. "We have premium quality tools" is a category claim every company makes. "We're free" is a price point, not a position, and the brand already refuses to lead with free. None of these survived a five-second test against an honest reader: so what?
The thing nobody else can copy is the operating model. Microapp is the first software company being built by one human and a team of AI agents — and we've written down how it works while it's still being built. Most company books are post-hoc: the founder retires, sits with a ghostwriter, reconstructs decisions from memory and selectively flattering archives. This one is being written with the agents, during the work, in commits the reader can pull and verify.
The single claim is the one paragraph I read myself when I'm about to ship something off-brand:
One human. An agent team. No investors, no consultants, no airline-shaped corporate scaffolding. Written down while it's still being built.
Six sentences. Three of them load-bearing. One human rules out the venture-backed founding team trope. An agent team names the operating model that didn't exist five years ago. Written down while it's still being built is the proof — every chapter of this book is anchored to a real commit, a real PR, a real artifact you can read.
The transferable why: the unique thing about your company is rarely the product. It's the way the product is made. Products get cloned in a quarter. Operating models take years. If you want a differentiator that survives competitors with bigger teams and better funding, lead with how you build, not what you build.
Locked 2026-05-06 · BRAND.md §0 · the operating story that runs through every chapter
That's the foreword. Three decisions: lead with the place, write for two tracks, brag with the operating model. Every chapter that follows is one of those decisions getting tested in the wild.