How we pick
The whole method, including what we refuse to claim. Every ranking page links here.
The questions that come before you have a shortlist. Whether one of these is worth buying at all, what the specification sheet actually means, and what the machine knows about your house.

Four questions come before choosing a model: whether a robot vacuum earns its place at all, what the numbers on the box actually measure, what the machine collects about your home, and on what basis anyone recommending one is making the call. This library answers those four.
The six hubs on this site are organized around floor problems, because that is how people actually shop. This library is the layer above that — the category-level decisions that apply whichever problem sent you here.
The honest answer to "are robot vacuums worth it" is conditional, and the conditions are specific. Floor type, floor plan, shedding animals, whether anyone in the house has mobility limits, and whether your floors are clear enough to run a robot without tidying first. Two of those are about the house and two are about the household, and they pull in different directions. Are robot vacuums worth it works through them without landing on a yes.
Pascals, millimeters of climb, millimeters of mop lift, minutes of runtime, liters of bin. Every one of those is published inconsistently and several are not standardized at all. Knowing which figures are comparable across brands and which are only comparable within one brand's range is the single most useful thing you can learn before shopping. The spec glossary covers each one and says plainly where the comparison breaks.
A modern robot vacuum builds a floor plan of your home and, on most current models, carries at least one camera. That data goes to a cloud service so the app can show you a map. Whether that matters to you is a personal judgment, but it should be an informed one, and the answer differs sharply by navigation type. What robot vacuums actually collect sets out what is known, what is not, and which models sidestep the question entirely.
We have not tested any of these machines. Nobody here owns one. What we do instead is compile the published specifications, record the source URL for every individual figure, mark the gaps as gaps, and rank on criteria stated before any product is named. That is a weaker claim than a lab makes and a stronger one than most affiliate sites can support. It is all set out on how we pick, and every ranking page on this site links to it.
Once you have a sense of the category, go to the hub that matches your actual problem: pet hair, carpet, hard floors, mopping, navigation and thresholds, or troubleshooting if you already own one and it is misbehaving. Every model we cover also has its own review page with its full published spec sheet.
If you have never owned one, start with whether they are worth it. If you are mid-shortlist, the spec glossary will save you the most time.
The whole method, including what we refuse to claim. Every ranking page links here.
A conditional answer rather than a yes, and the five conditions that decide it.
Every spec on the box, what it measures, and whether it is comparable across brands.
A floor plan, and often a camera. What each navigation type collects, and what does not.
Conditionally. They earn their place on hard floors and medium carpet in a home with clear floors, especially with a shedding animal. They earn it less in a small, tidy, mostly-carpeted flat where a stick vacuum does the job in ten minutes. The full conditions are on are robot vacuums worth it.
A cheap robot with mapping and an anti-tangle brush is a better buy than an expensive one with neither. Below about the price of a mapped model you are buying a bump-and-turn machine, which is genuinely useful for one room and frustrating across a floor.
No, and that is the wrong comparison. An upright cleans deeper in one pass. A robot cleans shallower, every day, without you. Frequency is the advantage, not power — which is why they suit homes that get dirty steadily rather than homes that need occasional deep cleans.
Many current models do — camera-based obstacle avoidance is standard on flagships and vSLAM navigation is camera-based by definition. LiDAR and dToF models navigate without one, though some still add a camera for obstacle recognition. The breakdown by model family is on privacy.
Mapped models effectively do, because the map, the schedule and the no-go zones live in the app. Bump-and-turn models like the eufy 11S MAX do not connect at all, which settles the privacy question by removing it.
We compile published specifications from manufacturer pages and listings, cite every individual figure, mark what is not published rather than filling it in, and rank on criteria stated before any product is named. We also name a pick to skip in every roundup. Full method on how we pick.
Every ranking on this site links back to how we pick. That is a trust requirement, not an SEO one.