Every page now arrives as a page, not as an empty shell

Until today this site was one application that drew itself in your browser after the fact. That works for a dashboard behind a login. It is a poor answer for a page somebody found through a search: what a search engine received was the same empty file for eighteen different addresses, with the same title and the same seventy-eight words, and whatever it made of that came later, if at all. The changelog you are reading was the clearest case, months of entries that were invisible to anybody who did not run JavaScript. Every marketing page is now built to finished HTML: the landing page went from nothing to just under three thousand words, this page to over four and a half thousand. The dashboard is untouched, it stays the application it should be.

• The landing page, the comparisons, the roadmap, this changelog, the security manifest, the legal pages, both forms and the founding page ship as HTML
• Each of them carries its own title, its own description and its own canonical address
• An address that does not exist now answers with a real 404 instead of a page that claims to exist
• One sitemap instead of two, generated from the pages that were actually built rather than maintained by hand
• The price on the landing page no longer waits on two hundred kilobytes of JavaScript before it appears
• Every internal link is checked against the built site before a release goes out, so a link to a page that was never generated fails the build