How this site works

What is here

39 self-cleaning litter boxes from 13 makers, priced from $100 to $899 on the makers’ own stores, each with a page of its own, plus 8 computed shortlists, 35 head-to-head comparisons and 7 explainers. The catalog was last refreshed on 2026-09-29.

A box is listed only while its maker’s own US store sells it new. Models that have left the store, pre-orders not yet shipping and semi-automatic sifters with no motor or sensor are left out, and the reasons are recorded alongside the data.

Where the figures come from

Every figure comes from the maker: its US product page, its user guide (several are published as PDFs) or its FAQ. Prices and consumable costs come from the maker’s own store. 71 source URLs are cited across the catalog, and every box page lists its own. Nothing is taken from a shop listing, and nothing is measured by us: there is no test bench, no decibel meter and no cats on staff.

Where a maker does not publish a figure for that exact model, the page leaves it out. A sibling model’s number is never borrowed, and where one maker page gives two different figures for the same thing (a few quote two noise levels, one quotes three kitten weights), the site either prints neither or takes the more cautious one, and says which.

What is worked out rather than quoted

The kitten gate. The lightest cat the maker says its sensor detects, converted to an age using the growth rule vets give owners of about a pound a month, rounded up, and never earlier than the maker’s own minimum age. 27 boxes publish a weight; 5 publish neither a weight nor an age and are labelled that way.

Days between emptying. Drawer litres ÷ (cats × 0.7 L a day), shown for one, two and three cats beside the maker’s own one-cat claim. Of the 20 boxes that publish both, 6 claim more than a quarter longer than the volume supports. Where the 0.7 L figure comes from.

Monthly running cost. Only for the 9 boxes that will not run without a proprietary consumable, and only where the maker publishes both a pack price and a usage rate (6 of them). Ordinary litter is left out, because every box needs some.

Big-cat fit. The maker’s stated maximum weight read against the 18 to 25 lb that large breeds commonly reach, with the entry opening beside it. A label, not a score.

Why there are no star ratings

A rating would be shown only with a citable source that is not a shop. None of the 39 boxes has one, so no stars appear anywhere and no review markup is sent to search engines. Nothing ranks boxes by an in-house score either: every list states the published figure it is ordered by.

Owner sentiment, and where it is missing

The pros and cons on a box page come from the maker’s figures, prices and documents. Anything attributed to owners sits in its own section under its own heading, and only for boxes whose owner discussion was actually read. Coverage today is 16 of 39; the other pages say so instead of showing an empty block.

How lists are ordered, and what this site does not do

The catalog’s default order is the kitten gate, lightest detectable cat first, because it is a published safety figure rather than a judgement. Brand and type pages are ordered by price. Every shortlist prints its rule and needs at least three qualifying boxes. Nothing is a paid placement, no maker has any say in what appears or in what order, and there are no shop links: no buy buttons, no tracking tags, no price widgets. Prices are approximate snapshots and will differ from what any shop charges today.

Safety notices

A recall, a CPSC notice, a maker safety update or a widely reported entrapment incident involving a listed model is printed on that model’s page, and on its card wherever it appears, with a link to the source. A recall also stops its structured data calling it available; a maker’s firmware safety update does not. 2 boxes carry one today, none of them a recall: no CPSC recall of a model in this catalog was found at the last update.

Who “we” is

A small independent project, not a publication. There is no editorial staff and no reviewer. Where the site recommends something it gives the checkable basis, such as “the lowest published detection weight in this catalog”, rather than an opinion from an implied expert.

Corrections are the fastest way to improve the data: send one.