Two site-only merges on Sunday. No app changes. Both came out of Rank's SEO pass.
Article pages now say who wrote them and when
super-notch #197 puts a byline under each article H1: By Austin Dixson · Updated <date>.
Nobody types the dates by hand. They come from git. A new script, scripts/page-dates.mjs, finds the last commit that changed a page's visible content. Nav, footer, CSS, analytics and schema-only commits don't count. It refuses to run on a shallow clone.
JSON-LD now adds an Article or BlogPosting with the author, a publisher reference, datePublished and dateModified. Sitemap lastmod follows each page's content date.
Some pages deliberately got no byline. Privacy, terms and refund keep only their Effective date. The Journal index is a Blog, and the help and release indexes are collection pages, with a publisher only. The dictate how-to stays publisher-only. Journal posts use their front-matter date as datePublished.
Buy, the GA4 snippet, the content security policy, titles, meta descriptions, H1s and body copy didn't change. The website suite ran 114 tests with 114 passing.
Three legal pages were naming a 404 as canonical
super-notch #198 fixed the canonical tags on privacy.html, terms.html and refund.html.
They pointed at https://supernotch.ai/<page>.html. Those apex deep paths return 404, because GoDaddy forwarding only redirects /. They now point at https://www.supernotch.ai/<page>.html.
Policy text, dates, links, checkout and the sitemap are unchanged.
Check where your canonical actually lands
A canonical tag tells search engines which URL to index. If that URL 404s, the tag sends them to a dead page. Loading each canonical URL once takes a minute.
Built with Cursor cloud agents and Node's test runner.