Buying guides
Home-fit finder
Buying guide
Verdict
Swept's pick for a robot vacuum that avoids dog poop is the eufy Omni S2, for mostly hard-floor homes with rugs and a puppy or a pet that has accidents.
BuyBest overall for puppies, cords and toys
Opens Amazon.com · Prices change often
At a glance
SuitsMostly hard-floor homes with rugs and a puppy still learning, or a pet that has the odd accident
The catchLeaves hair pressed into carpet behind, and its mop dampens rug edges
SuitsMostly carpeted homes with a dog that sheds hard and has the odd accident indoors
The catchIts aim is surer on bare floor than on a rug
SuitsWood, LVP or tile homes with thin rugs, an ordinary shedder and the occasional accident
The catchOnly one lab has tested its eyes, and thick rugs can trap it
SuitsMostly hard-floor homes on a tight budget where the robot has to see what's on the floor
The catchPoor at pulling pet hair out of carpet
Wait and Skip
Why you see a price band, not a price Prices change daily, so our pages show only the price band a robot sits in and leave the current price to Amazon. That way a page is never wrong by Tuesday.
Robot makers now promise to recognize hundreds of household objects. In a puppy home, only one object matters. It's the one nobody enjoys planting on a test floor, and very few robots steer around it every time. Swept's pick is the eufy Omni S2. It scored a perfect 24 of 24 on Vacuum Wars' obstacle course. RTINGS watched it avoid fake pet waste on every run, on carpet and bare floor alike. Both labs bought their own robots, and no robot here has a better record. Buy it for mostly hard floors with rugs and a puppy still learning, or an older dog who doesn't always make it to the door.
The catch is carpet. It lifts loose hair well enough, but hair trodden into a rug tends to stay there. Its roller mop also can't rise far enough to keep rug edges dry. If most of your house is carpet, the Dreame X60 Max Ultra Complete below is the better robot.
Mostly carpet? The eufy still sees on rugs, but it cleans them only adequately. The Dreame X60 Max Ultra Complete cleans carpet far better and still spots most messes.
Hard floors, but with deep rugs and a dog that sheds hard? That home leans toward the X60 too. Its deep-rug cleaning and reach into edges make it our pick there in best robot vacuum and mop. The eufy remains the choice when your rugs are thin and the mess on the floor is the bigger worry than the hair.
Money tight? The MOVA P10 Pro Ultra has a real camera and a decent poop record for much less. For a superb mop, the ECOVACS T90 PRO Omni is the step up.
Worried about a camera in the house? Know the trade first. Every robot that reliably avoids dog poop does it with a camera. The eufy is the gentlest version. Its US guide lists no microphone, and Vacuum Wars, testing a robot it bought, found the app offers no live video. If a camera is a hard no, our camera-free picks are good robots. You'll just be picking up before every run.
A dog that sheds like it's paid to? Avoidance and hair pickup are different jobs, and our pet hair picks answer the second. Still unsure? The Home-Fit Finder sorts it by floor, pet and budget.
BuyBest overall for puppies, cords and toys
Price band: under $1,600
Eufy says its 3D camera and depth sensor recognize more than 200 kinds of object. For once, a box claim survives the test course. Vacuum Wars' 24-object course came back perfect, a score only one other robot has managed there. RTINGS found it avoided pet waste on every run, on carpet and wood, and rarely needed rescuing. The Hook Up did watch it roll over its fake poop. That was one object on one run, though, and we weigh the bigger samples.
At home, that means you leave for work with a puppy in the house and come back to clean floors. No brown stripe down the hallway. That's the whole job, and the S2 does it.
The one honest catch is carpet. Loose hair and grit come up fine: Vacuum Wars measured 87% of embedded sand out of medium pile. Hair pressed into a low rug is another story, and RTINGS saw a lot of it left behind on default settings. On a mixed floor, the roller mop lifts but has no cover, so rug edges come back damp. Eufy's own product FAQ suggests no-go zones for rugs thicker than 2 in (50 mm).
For mostly hard floors and a dog still learning the rules, we'd buy it. The full eufy Omni S2 review covers the mop, the dock and the upkeep.
BuyBest for carpet-heavy homes
Price band: under $1,700
Dreame fits two AI cameras and promises 280 object types. The labs mostly agree it can see. RTINGS found it avoids pet waste reliably, and Vacuum Wars scored it 22 of 24. A 30-day reviewer watched it clear two fake messes on carpet without touching them.
Where it pulls ahead is the carpet itself. Vacuum Wars measured 89% of embedded sand from medium pile, plus every strand of its flattened pet hair. Vacuum Wars saw it clear a 2 in (51 mm) step, and The Hook Up saw it cross a step of about that height in both directions, so the sunken den is no obstacle. Dreame backs it for three years, the sort of promise we like to see in writing.
The catch: its aim is surer on wood than on a rug. RTINGS said as much, and Vacuum Chef saw it brush a fake mess on carpet. Better than nearly everything, not flawless.
For a carpeted house with a shedding dog who slips up now and then, we'd buy it. The Dreame X60 Max Ultra Complete review has the rest.
BuyBest mopper that still watches the floor
Price band: under $600
ECOVACS pairs a camera with structured light and calls it AIVI 3D 4.0. Vacuum Wars scored it 23 of 24, one object short of perfect. Its roller mop is the best of these picks. Vacuum Wars scored its dried-stain cleanup at 157, against that lab's average of 108.
The catch is that its eyes rest on one lab. Nobody has published a dedicated pet-waste run on this exact robot. On thick rugs, owners report it can jam. It belongs on hard floors with thin rugs.
If you want floors mopped as well as spared, we'd buy it. Just know its poop record is shorter than the eufy's. The T90 PRO Omni review covers the mop in detail.
BuyBest on a tight budget
Price band: under $500
At this price, most robots find dog mess the way a bowling ball finds pins. The P10 Pro Ultra has a real camera. RTINGS found it won't attempt to clean pet waste, even on carpet. Vacuum Nerds saw it give fake poop a wide berth, and Vacuum Wars scored it 19 of 24. One tester disagrees: TechGearLab, testing in a home, saw it plow through pet messes. Three for, one against. Good, but well short of the eufy's record.
Buy the original. The newer P10 Pro Ultra Gen 2 drops the camera, a sequel that removes the one thing the first film had going for it.
The catch is pet hair in carpet. Vacuum Wars measured 66% of its flattened hair, and TechGearLab 40%. On hard floors it's fine.
For mostly hard floors on a budget, we'd buy it.
WaitEqual-best eyes, but wait
Price band: under $1,600
The Saros 10R is the only other robot with a perfect 24 of 24 at Vacuum Wars. RTINGS found it keeps a careful distance from pet waste. For poop, its eyes match the eufy's. USB cords are its weak spot: RTINGS saw it misread them and sometimes tangle.
The reason to wait is hair, where the labs can't agree. With two runs at Vacuum Wars, it lifted 92% of the flattened pet hair. With real fur worked into carpet at TechGearLab, it managed 56%. Its two-piece brush tends to bunch hair where the halves meet, so some of it never reaches the bin. Long-haired dog owners keep reporting a bin that clogs well before it's full. A current-firmware test that weighs both bin and brush would settle it, and might change our mind.
We'd wait, and buy the eufy Omni S2 today. The Saros 10R review has the full split.
The eufy S2 and the Dreame X60 are both 2026 robots with solid records. Two newer ones aren't there yet.
The MOVA V70 Ultra Complete went on sale in August. Vacuum Wars scored it 23 of 24, and Just A Dad Approved calls its avoidance genuinely excellent. That's one US lab and one reviewer, with no dedicated pet-waste test. An owner of two weeks says charging cables still trip it. Its bagless dock is also unproven with heavy shedders. One owner with several shedding dogs found hair piling up at the dock's intake and the dock filter caked within a week. Wait.
The iRobot Roomba Max 715 Vac arrived in May. iRobot says PrecisionVision AI steers it around pet waste and cords. The only test so far is a three-month home trial. It threaded through clutter cleanly. Its dock, though, didn't once manage to suck the robot's bin clear in five attempts, which rather undermines the point of a self-emptying robot. Its predecessor's lab record doesn't carry over to a new robot. Wait for a lab.
Neither is worth holding out for. The eufy has the record now.
We read every obstacle-course and pet-waste result we could find for each exact model. Then we set them against what each maker promised. A robot made the picks only if independent testers had run that exact model through an obstacle test, and fake-waste results counted most. The T90 PRO Omni is the one pick resting on a general obstacle course alone, and its section says so. This page weighs 20 models, and 17 of them came with their maker's manual. The site's research as a whole rests on 497 sources we kept in full, plus 148 Reddit threads we read and kept only as summaries. The test-by-test record sits in the evidence section below.
Yes, if it has a camera. The two best passed every large test. What nobody has tested is your dog. Every lab uses a fake: a firm prop on a clean floor in good light. A fresh, soft accident on a patterned rug at 6 a.m. is untested by anyone. Treat a great score as the odds tilted firmly your way, not a guarantee.
Because it ran over the fake poop in every dedicated pet-waste test. RTINGS concluded it isn't dependable around pet waste. The Hook Up's robot drove straight over its fake mess and scored 3 of 6. George Langabeer tried two Saros 20s. Roborock blamed beta firmware for the first, yet neither ever avoided the fake poop. Tom's Guide calls its avoidance unbeatable. For a home with accidents, that would put the wrong robot on your floor.
Its 17 of 24 at Vacuum Wars is respectable, but that course is mostly everyday clutter. Roborock's page says it recognizes more than 300 kinds of object. On every pet-waste test, fake dog poop wasn't reliably one of them. In aerospace manufacturing I learned to treat a capability with no test behind it as a parts list. Here the tests exist, and the answer is no. RTINGS thinks it's a software problem, so firmware may fix it. No one has retested.
What separates you from the readers who should buy one is the accident. With house-trained pets, it remains our choice for steps between rooms in several places. See our home page and Saros 20 vs Dreame X60. It's also in our best robot vacuum for carpet, best robot vacuum for long hair and best robot vacuum without a mop roundups. Our guides to whether robot vacuums are worth it and whether robot vacuums are banned pick it too. The Saros 20 review has the detail.
If your dog has accidents, skip the Saros 20. If you also have a step to climb, buy the Dreame X60 Max Ultra Complete. It climbs 2 in (51 mm) and sees the mess.
Yes. Lidar sees walls and furniture, not a mess. Vacuum Wars scored the camera-free Dreame D20 Plus 0 of 24 and the Roborock Qrevo S Pro 5 of 24. The camera-free eufy Omni C28 ran over fake poop in a three-month home test. For homes that pick up first, good camera-free robots exist, like the Dreame D20 Plus, the Roborock Q7 M5+ and the Tapo RV30 Max Plus. For a puppy home, they're the wrong tool.
| Robot | Best for | Pet waste, by the testers | Carpet | The catch | Amazon |
|---|---|---|---|---|---|
| eufy Omni S2 | Puppies and accidents on mostly hard floors | Perfect at two labs; one single-run miss | Ordinary on pressed-in hair | Damp rug edges when mopping | View on Amazon (opens in a new tab) |
| Dreame X60 Max Ultra Complete | Carpet-heavy homes | Reliable at RTINGS; brushed it on carpet at one lab | The best here | Aim slips a little on rugs | View on Amazon (opens in a new tab) |
| ECOVACS DEEBOT T90 PRO Omni | Hard floors and a top mop | One lab's course, near perfect | Thin rugs only | Thin poop record | View on Amazon (opens in a new tab) |
| MOVA P10 Pro Ultra | Tight budgets | Three testers for, one against | Weak on pet hair | Buy the original, not Gen 2 | View on Amazon (opens in a new tab) |
BuyBest overall for puppies, cords and toys
Other robots clean carpet better, and one or two cost less. None has a cleaner record of seeing the mess and going around it. In a puppy home, that's the promise that counts. For mostly hard floors and a dog still learning, buy the S2.
| Robot | Vacuum Wars 24-object course | RTINGS pet waste | The Hook Up poop test | Other testers | Call here |
|---|---|---|---|---|---|
| eufy Omni S2 | 24 of 24 | Avoided every time; rarely needed rescuing | Ran over the fake poop (one run) | None on pet waste | Buy |
| Roborock Saros 10R | 24 of 24 (max sensitivity) | Consistent, keeps its distance; often misreads USB cords | Not tested | TechGearLab: avoided over 90% of obstacles; owners split on cables catching the mop pads | Wait |
| Dreame X60 Max Ultra Complete | 22 of 24 | Reliable; better on bare floor than carpet | 4 of 6; ran over the poop once at the start, then avoided it | Jamie Andrews: cleared two fake messes on carpet and a dangling cord. Vacuum Chef: 88%, perfect on hard floor, touched fake poop and a USB cable on carpet | Buy |
| ECOVACS DEEBOT T90 PRO Omni | 23 of 24 | Not reviewed | Not tested | Sponsored three-month owner: avoids cables and pets; one-week owner: sees toys and dogs but dithers | Buy |
| MOVA P10 Pro Ultra | 19 of 24 | Won't attempt pet waste, carpet included | Not tested | Vacuum Nerds: avoided fake poop, shoes, toys and cables. TechGearLab: often plowed through pet messes | Buy |
| Roborock Saros 20 | 17 of 24 | Not dependable around pet waste | 3 of 6; ran over the fake poop | Langabeer: neither of two robots ever avoided fake poop. Tom's Guide: "unbeatable". ModernDayTech: avoidance stayed great, but seemed off on the way back to the dock | Skip; buy the X60 for accidents plus steps |
| MOVA V70 Ultra Complete | 23 of 24 | Not reviewed | Not tested | Just A Dad Approved: genuinely excellent | Wait (new) |
| Roomba Max 715 Vac | Not tested | Not reviewed | Not tested | Just A Dad Approved: worked around its clutter course cleanly | Wait (new) |
| Roomba Max 705 Vac | Not tested | Result behind RTINGS' paywall | Not tested | Just A Dad Approved: avoided a cable, a toy and fake poop on hardwood | Considered |
| Roborock Qrevo Curv 2 Flow | 16 of 24 | Not consistent | 2 of 6; ran over the fake poop repeatedly | None | Considered |
| Matic | 14 of 24 | Consistently avoided pet waste, but came close; beta pet-waste feature was off | Not tested | Owners: mostly avoids cables, socks and toys | Considered |
| eufy Omni C28 (no camera) | 15 of 24 | Not reviewed | Not tested | Just A Dad Approved: ran over fake poop and cables | Considered |
| Dreame L40 Ultra Gen 2 (no camera) | 2 of 24 | Not reviewed | Not tested | Vacuum Chef: 30%, failed on cables and fake pet waste | Wait (new this year), not for this buyer |
| Roborock Qrevo S Pro (no camera) | 5 of 24 | Not reviewed | Not tested | Just A Dad Approved: ran over a cable and toys | Considered |
| Dreame D20 Plus (no camera) | 0 of 24 | Not reviewed | Not tested | None | Considered |
| Roborock Q7 M5+ (no camera) | No front obstacle sensors | Not reviewed on this model | Not tested | Owners: gets stuck on furniture bases | Considered |
| Tapo RV30 Max Plus (no camera) | Not tested | Not reviewed | Not tested | TechGearLab: runs over socks and cords | Considered |
| Roomba 105 Vac (no camera) | Not tested on this model | Not reviewed on this model | Not tested | Just A Dad Approved: practically nil, runs over cables and toys | Considered |
| Shark AI Ultra AV2501S (no camera) | Not tested | No camera or object recognition | Not tested | Just A Dad Approved: advertised avoidance didn't hold up | Considered |
| Narwal Freo X10 Pro (no camera) | Not tested on this model | Not reviewed | Not tested | US manual lists no camera | Considered |
| Tester | How it tested | What to bear in mind |
|---|---|---|
| Vacuum Wars | Buys its robots; 24-object course at maximum avoidance sensitivity | The course is mostly household clutter, not pet waste |
| RTINGS | Buys US retail robots; room half carpet, half hard floor, with cords, socks, pet bowls and fake waste; filmed over three runs | Its X60 run used firmware 4.3.9_1495 |
| The Hook Up | April 2026, unsponsored; one fake poop among a few objects, single run | One miss weighs heavily in so small a sample |
| George Langabeer | Two Saros 20s sent by Roborock, April and May 2026 | Roborock blamed beta firmware for the first; the second fixed cords but not poop |
| Just A Dad Approved | Months-long home tests with a dog | Home conditions, not a lab |
| Jamie Andrews | 30 days with the X60 | One reviewer's home |
| Vacuum Chef | X60 and L40 sent by Dreame; says neither review was sponsored | Maker-supplied robots |
| TechGearLab | Tests in homes | Different props and lighting from a lab |
| Vacuum Nerds | P10 Pro Ultra original against Gen 2 | Robot's market not stated |
| Robot | Detail moved here from the picks above |
|---|---|
| eufy Omni S2 | RTINGS scatters socks, cords, pet bowls and fake waste across a room half carpet and half hardwood, and films each robot three times; the S2 avoided pet waste on every run. Vacuum Wars and The Hook Up both found no live camera view in the app |
| Dreame X60 Max Ultra Complete | The 30-day reviewer also saw it avoid a dangling white cord; The Hook Up saw it drive over a fake mess on its first pass, then avoid it after that |
| ECOVACS DEEBOT T90 PRO Omni | A three-month owner, sponsored by ECOVACS, says it steers around cables and pets |
| MOVA P10 Pro Ultra | RTINGS found its obstacle handling very good overall; Vacuum Nerds also saw it avoid a shoe, a cat toy and two cables |
| MOVA V70 Ultra Complete | The two-week owner also reports trouble with thin, low objects. The heavy-shedder owner, on this exact model for one to two weeks, says the robot itself stayed clear of wraps and clogs; the hair gathered at the dock's suction port and pipe instead, and the bagless tank filled faster than their old V50's bag. That is one owner's report: no lab has measured how this model's bagless dock copes with heavy shedding |
| Source | What it found | Was pet waste the test? |
|---|---|---|
| Tom's Guide | Called its avoidance "unbeatable" | Method not stated |
| ModernDayTech | Avoidance stayed great over three months (media sample) | General clutter |
| Vacuum Wars | 17 of 24, middle of the field | Mostly clutter |
| RTINGS | Not dependable around pet waste | Yes, failed |
| The Hook Up | Ran over the fake poop | Yes, failed |
| Langabeer, first robot | Never avoided the fake poop | Yes, failed |
| Langabeer, second robot | Never avoided the fake poop | Yes, failed |
Every source that set out to test pet waste saw it fail. RTINGS puts the drop from the 10R down to software, so a future update could change this, but no retest on late-2026 firmware exists. For a home with accidents, skip the Saros 20; if you also have a step to climb, the Dreame X60 Max Ultra Complete is the robot to buy.
| Robot | Why it lost here |
|---|---|
| Roborock Saros 20 | Fails pet waste (above); with house-trained pets it remains our pick for steps between rooms |
| Roborock Qrevo Curv 2 Flow | Ran over fake poop repeatedly; a fine mopper where accidents never happen |
| Matic | RTINGS saw it keep clear of fake pet waste, but also saw it struggle to navigate on every surface, and at 7.8 in tall it can't get under low furniture. Discussed only |
| Roomba Max 705 Vac | One home test avoided fake poop and Consumer Reports found its avoidance better than many robots', but RTINGS' pet-waste result is paywalled. In that home test the dock failed to clear the bin on its first try after carpet, and owners of a 705 of unstated variant report auto-empty trouble too |
| Camera-free robots (D20 Plus, Q7 M5+, Tapo RV30 Max Plus, Omni C28, Qrevo S Pro, Freo X10 Pro, Roomba 105 Vac, Shark AV2501S) | None can see a mess; several are good buys elsewhere on this site for homes that pick up first |
| Dreame L40 Ultra Gen 2 | New this year and camera-free; early tests saw it hit almost everything. Don't confuse it with the first-generation L40 Ultra, which had a camera |
| Question | What the sources found | Why the call stands |
|---|---|---|
| eufy S2 on carpet | The Hook Up measured 78.8% of a mixed debris load from high pile at maximum suction over two passes; Vacuum Wars measured 87% of embedded sand and 88.5% of flattened pet hair; RTINGS, on Standard suction in one pass, saw much of its hair left pressed into low pile | Loose material comes up, pressed-in hair doesn't; the S2 is picked for hard floors, and the X60 covers carpet |
| Saros 10R pet hair | 92% at Vacuum Wars over two runs; 56% at TechGearLab with real fur worked into carpet, passes and setting not stated | Passes and hair load explain most of it; it's why we wait |
| X60 pet hair | Every strand of flattened hair at Vacuum Wars; only passable at RTINGS in one default pass | Its avoidance, not its hair pickup, earns its place here |
| P10 Pro Ultra pet waste | RTINGS and Vacuum Nerds saw it avoid fake waste; TechGearLab saw it plow through | Likely different props and lighting, lab against home; the majority holds |
| Pre-launch eufy S2s | Robots from January 2026 were taken back for a hardware upgrade | Every lab result above is on robots sold since the March launch, except Vacuum Chef's possibly early S2, which isn't used for the call |
| X60 thresholds | Vacuum Wars: cleared a 2 in (51 mm) single step; The Hook Up: crossed a step of about that height in both directions; RTINGS: its test tops out lower, and tall thresholds must be marked on the map first | Method and settings explain the spread; the climb is not why it's picked |