Publii is absolutely amazing.
I wrote my own CMS using OO design at its core so that any thing within the CMS could be referenced to as an object. Everything was stored in MYSql and could be sliced, diced, and julianmed. The core of the CMS was the hierarchy object. This object, similar to a computer directory in reference, parent/child, could organize HTML elements, code, snippets, template fragments, Schema JSON, data, and references including aliases in wild and weird ways as objects at any point. It used a templating theme similar to Handlebars as references or references to references that was stored within the database and could do incredible things. I could replicate the entire DMOZ site exactly as they had it using only 6 characters of content. That is it. Truly a marvel.
So why did I tell you this?
My CMS was designed for automation, though not a requirement. I used it to reference data within about 170 tables with many millions of rows to produce nearly 24 million pages of hack/attack data for research. Each page was rendered upon request based upon this OO design. It is lightning fast.
My point is this.
Publii, with a rather simple API, could allow coders to generate pages, render, and deploy sites outside of the GUI. I can see a use for this with some ideas I have floating around. Would creating an API to Publii be a consideration?
So what about my CMS?
As radically cool as it is, no-one has made another in the decade plus since I began coding it, Publii makes sites awesomely simple and therefore preferred for my remaining sites. They are relatively simple and could easily become more blog like. I am busy making new custom themes to make this happen. Once I deploy even one site, I will be donating.