Context

Currently, creating a full-fledged website already takes about a day, which is radically cheaper than previous development, but against the background of modern AI-builders, this is still a long time. The goal is to figure out whether the infrastructural part of the launch can be reduced to such an extent that creating dozens of new websites stops being organizationally expensive.

Problem

A significant part of the time may be spent not on the unique logic of the website, but on repetitive actions: repository, runtime, environment, database if necessary, domain/SSL, deploy, basic observability, backups, and agent connection.

Direction

It is necessary to measure the actual current launch path of several websites, find repetitive manual steps, and eliminate the most expensive ones one by one.

One should not build a huge universal scaffolding in advance. In accordance with the general philosophy of the project, it is more useful to automate specifically detected repetitions and keep the foundation understandable.

Connected with runtime, environments, deploy, observability, and the existing task of minimal project management.