How We Work

Last updated October 4, 2026.

These are the rules every page on this site follows, and what AI does and does not do here.

How we decide what gets shown on this site

• Every specific claim links to a real source we've read in full, not a search snippet, not someone else's summary of it. If we can't get the whole thing, we don't publish it.
• A known issue doesn't require a single shared firmware version across all models. Mammotion's model lines plausibly share upstream code with different per-model release numbering, so the same real bug can surface at different version strings. The corroboration bar depends on how many independent reports we have: one report stays behind the scenes entirely; two or more independent reports of the same specific symptom, without a confirmed firmware version, appear on issues we're watching; three or more (or two or more with a confirmed firmware version) get a full guide page.
• No usernames or real names appear in body copy, on either side of a report. Identity is reachable only by following a citation link to the actual source.
• Facts are stated plainly, without editorializing, regardless of whether they flatter or embarrass a manufacturer.

This is a summary of our full standard: a much longer and more detailed, versioned internal document we don't publish, but every rule above traces back to it. How the pages themselves get written, including what AI does here and what it is not allowed to do, is below under how this site is made. Think something here is wrong or out of date? Request a correction. That page also covers what happens next.

What “checked” means here

We don't run lab tests, and we don't own every model. We read what owners and manufacturers publish, and a fact goes on a page only when it links to a source we read in full. Owner reports are counted against the bar above, so one report alone never becomes a guide. Sources in other languages are read through an AI translation, so from them we use only plain facts that survive it: a version, an error code, a date, what happened.

How this site is made

I research and write these pages with AI assistance and a (future) team of volunteer human reviewers. A custom built AI tool helps me find, read, translate and draft from many sources, but it never makes the decision to publish, because that decision belongs to me alone.

Every fact is human reviewed by me and my team against a source that is linked on the same page, so you can check it too. A search snippet, someone else's summary, or something an AI told me never gets onto this site as a fact. When we get something wrong I correct it in place and log it on corrections.

Any new page or update cannot go live until a person has read it and I have decided it is ready. This editorial review process can delay 'hot topics' from immediately appearing on site, but that delay is a tradeoff I'm willing to make. Accurate information always beats fast noise.

What I do with your information, which is deliberately little, is in the privacy policy.

Who runs this site

Michael runs it. He's a robot mower owner in Tennessee, and the site is independent of every brand. The about page has the story.

We have no affiliation with any brand covered here. Limited advertising offsets what it costs to build and run the site. No advertiser decides what we publish.

What we don't do

• No sponsored placement and no paid ranking.
• No claim without a linked source.
• No AI deciding what gets published. A person reads or edits every page first.
• No names of owners in our text.
• No silent corrections. When we get a fact wrong, the fix is dated and listed on corrections.