Design of a webapps factory as a separate project
Design work on a factory for small, niche web apps. Two conclusions: it has to be a separate product from our client platform, and the slow part is not the code.

We have been designing what it would take to launch and run small web apps at a much faster pace — each built around a single need for a specific audience, rather than another general-purpose tool. This round was design work, not code, and it produced two conclusions: the factory has to be a separate product from the platform we run client sites on, and the slow part is not building the apps.
It has to be its own product. Our current platform exists to run websites for clients: each site is designed for one company, approved by them, and handed over with an admin panel they own. A factory of niche apps works the other way round — no client to sign off a design, no per-app admin panel, a shared design kit instead of bespoke work every time. Serving both from one system would make both worse, so the factory will be a separate project with its own resources.
The friction isn't where we assumed. We expected building each app to be the slow part. It isn't — most of that can be generated in hours. The slow parts touch the outside world: registering domains, opening social accounts, setting up payments. None of them compress well per app, so they get set up once, up front, and every new app draws on them. Apps will launch on a subdomain and earn their own domain only once they show real traction.
Meanwhile our platform stays what it is: the system we run client projects on, and where we prove a technique before reusing it. We're still shaping the final form.
Keep reading
- Capital Is Getting Expensive, Intelligence Is Getting CheapThe cost of capital is rising while the cost of intelligence is collapsing. Their crossing redefines how companies should fund growth.
- The Case for AI-Native CompaniesAI-native companies produce roughly 100x the output of legacy firms that adopt AI on top of old hierarchies. Here is why organizational design, not software, is the deciding factor.
- GTM Engineering and Growth Engineering Are Not the Same RoleGTM engineering automates revenue systems outside the product. Growth engineering runs experiments inside it. Here is where each discipline starts and stops.
- Countermeasures against prompt injection in our AI agentOur AI agents are being probed for their system prompts. The refusal holds, but nothing was detecting, logging or throttling those attempts.
