Fifteen articles in this archive were dated before the site existed

Fifteen of the seventeen articles in this archive carried a publication date from before app-builder-index.com existed. The earliest read May 14, 2026 on a domain that was not registered until August 19, 2026, a gap of ninety-eight days. Those dates were visible in the byline on every affected page and in the machine readable structured data underneath it. They are corrected as of today.

What was wrong

When this site was seeded, its launch archive was written with publication dates spread backwards across the preceding months. The intent was presentational, to make an index that opens with fifteen buyer guides read like a publication with a back catalogue rather than a directory that appeared overnight. The effect was a false claim, repeated fifteen times, in the one field search engines read to decide how old a page is.

The registration record settles it. app-builder-index.com was registered on August 19, 2026 at 10:10 UTC. Anyone can check that in about ten seconds. No article on this domain can honestly carry a date before it.

A second field was wrong in the same way, and we did not notice it until we went looking. The last updated date, also published in structured data, predated the domain on thirteen of those fifteen pages.

What we changed

Every affected article now carries a publication date on or after August 19, 2026, spread across the days the archive was actually written rather than collapsed onto a single timestamp. Collapsing them would have put thirteen long-form guides at one identical second, which is a worse signal than the one we were fixing. The article text is untouched in every case, and each affected page carries a dated note at its foot saying what happened to its date.

Two pieces moved further than the rest, because each contains its own dated update that has to come after its publication. Our MCP buying criterion piece and our index recompute piece are both now dated August 20, 2026, ahead of the updates they carry.

What we did not change

Nothing about the ratings, the axis scores, the weights or the rankings moved. This was a date correction and only a date correction.

We also did not flatten the ordinary practice of stamping an article at a tidy slot time a few hours either side of when it was finished. That is normal editorial practice and it is not what was wrong here.

The check that would have caught it

Compare every published date against the registration date of the domain it sits on. That is one lookup and one comparison, and until today nobody here was running it. It now runs on this site on a schedule, alongside the check for articles dated in the future, which we were already running and which had never found anything. Looking forward for impossible dates and never looking backward is how ninety-eight days of imaginary archive sat in plain sight.

We publish our weights so readers can catch us being wrong about scores. This one was not about scores, and it should not have needed an internal audit to find.