Low-Profile Robot Vacuums: What Fits Under Furniture

That matters because missed under-furniture dust is one of the quiet reasons buyers become disappointed. A robot can map the house perfectly and still leave gray lines under the sofa if its sensor tower is too tall.

The short version: if your furniture has only 8 cm of usable clearance, even the best low-profile flagships are still tight. If you have 9-10 cm, the new class of embedded, solid-state, or retractable LiDAR robots becomes much more useful. If you have 11-12 cm, height stops being the main problem and obstacle avoidance, dock size, carpet behavior, and maintenance cost matter more.

Which low-profile robot vacuums fit under common furniture?

Here are the most relevant models in ui44's database right now, sorted by the height number a buyer should check first.

The important pattern is not that one brand wins every room. It is that the old "LiDAR tower equals premium navigation" rule is breaking. Roborock's Saros 20 uses embedded 3D ToF LiDAR. Dreame and MOVA use retractable sensor modules. Ecovacs has already shown that embedded dToF can sit in a 9.8 cm body. Bosch goes another direction and makes the whole appliance disappear into cabinetry.

That is a real smart-hardware shift, not just a cleaning spec race.

Why does the LiDAR tower matter so much?

Traditional robot vacuums often put a spinning laser rangefinder in a raised puck on top of the robot. That is a great place to see the room, but a terrible place to fit under a bed frame. The tower becomes the highest point, so a robot with a normal 8-9 cm body can effectively become an 11-12 cm robot once the sensor is included.

The newer designs attack the problem three ways.

Embedded LiDAR or 3D ToF keeps the body consistently low. The Roborock Saros 20 is the cleanest example in the database: roughly 7.98 cm tall, 36,000 Pa suction, StarSight 2.0 with embedded 3D ToF LiDAR, VertiBeam lateral avoidance, and recognition for more than 300 object types. RTINGS also singled out the Saros 10R family as a strong example of solid-state LiDAR and obstacle recognition without a tall profile. The tradeoff is that buyers should look carefully at reviews for mapping stability around black furniture, mirrors, floor-length curtains, and chair legs.

Retractable LiDAR gives you two heights. Dreame X60 Max Ultra Complete is 79.5 mm with its VersaLift DToF sensor retracted and 102.8 mm when raised. MOVA Mobius 60 lists 89.5 mm retracted and 111.5 mm raised. This design makes sense in real homes: the robot can lower itself under furniture, then raise the sensor in open spaces. The question to ask is whether obstacle detection, mapping confidence, or cleaning behavior changes when the module is down.

Camera-heavy systems add semantic understanding. Dyson's Spot+Scrub Ai is not especially short at 110 mm, but its AI-powered camera can identify nearly 200 household substances and stains, then adjust cleaning passes. Dreame's X60 and MOVA's Mobius 60 combine retractable depth sensing with AI cameras. That is useful if the space under the sofa also contains charging cables, toys, pet bowls, or slippers. It also means privacy settings and image-processing policies deserve more attention.

What should you measure before buying?

Do not measure the front lip of the sofa and call it done. The lowest part is often a rear rail, a center support, a caster bracket, or a sagging fabric panel. A robot that enters from the front can still wedge itself deeper inside.

Use this simple rule:

Furniture gap should be at least the robot's operating height plus 1 cm.

That margin matters because floors are not perfectly flat, rugs add height, upholstery sags, and the robot may pitch slightly while crossing a threshold. If your sofa gap is exactly 8.0 cm, a 7.98 cm robot is technically short enough but practically risky. A 9.0 cm gap is much safer.

For retractable designs, write down both heights:

• The retracted height tells you whether it can physically enter.
• The raised height tells you how tall it becomes in open mapping mode.
• The behavior under furniture tells you whether the robot still navigates intelligently when the sensor is down.

That third point is the one spec sheets often hide. A robot that lowers itself but becomes less confident under a sofa may still clean, but it may also slow down, miss edges, or avoid the area after a few bumps.

How do the top low-profile designs compare?

The three most interesting low-profile approaches right now are Roborock's always-low Saros design, Dreame's retractable VersaLift design, and Bosch's built-in kitchen approach.

Roborock Saros Z70 and Saros 20: thin body, premium feature set

The Saros Z70 is 7.98 cm tall, includes StarSight 2.0 LiDAR, 3D structured light, an RGB camera, 22,000 Pa suction, and Roborock's OmniGrip five-axis mechanical arm. The arm is not a reason to buy a vacuum by itself, but it is relevant to low-clearance cleaning because socks, small toys, and shoes are exactly the objects that block a robot from reaching under furniture.

The Saros 20 is the more pure cleaning-spec version: roughly the same low body height, 36,000 Pa suction, embedded 3D ToF LiDAR, object recognition for 300+ object types, and AdaptiLift Chassis 3.0 for thresholds up to about 8.8 cm. That makes it a better fit if your problem is not just getting under beds, but moving between thick rugs, threshold strips, and hard floors without needing rescue.

The caution: ultra-slim flagships can be expensive, and not every home needs the full spec sheet. If your sofa clearance is 11 cm and your floors are mostly hardwood, the low-profile premium may be less important than reliable mopping, simple maintenance, or a smaller dock.

Dreame X60 and MOVA Mobius 60: retractable sensors plus obstacle AI

The Dreame X60 Max Ultra Complete is one of the clearest examples of a two-height strategy. It lists a 79.5 mm body when the sensor is retracted, 102.8 mm when raised, 35,000 Pa suction, dual AI cameras, lateral 3D structured light, and recognition for more than 280 object types. It also claims double-layer threshold crossing up to 8.8 cm.

MOVA Mobius 60 takes a similar path with FlexScope retractable DToF LiDAR: 89.5 mm retracted, 111.5 mm raised, 30,000 Pa suction, AI SmartSight, structured-light sensing, and claimed recognition for more than 240 object types. Its separate smart angle is MopSwap, which automatically selects among three mop pads by room type.

These retractable systems are attractive if your home is mixed: low sofas in one room, thick thresholds elsewhere, and plenty of loose objects. The raised mode can help in open areas; the lowered mode can reach where a fixed tower cannot. The buyer question is reliability: will the robot consistently remember those under-sofa zones, or does it treat them as risky spaces to avoid?

Bosch BCRI3BX1: the kitchen-plinth answer

The Bosch BCRI3BX1 is easy to overlook because it is not a normal freestanding flagship. It is a built-in vacuum-and-mop robot designed to live inside a kitchen plinth, with a connected full-service station for dust emptying, mop washing at 70 °C, 45 °C hot-air drying, and station self-cleaning.

At 8.4 cm tall, it is relevant to low-profile buyers for a different reason: kitchens have toe kicks, cabinet bases, and appliance edges that standard docked robot vacuums often handle awkwardly. Bosch is saying the robot should be part of the room architecture. That will not help renters or anyone who wants one robot for the whole home, but it is a useful signal for where premium floor care is heading: lower, more integrated, and less visually intrusive.

Is a low-profile robot vacuum always better?

No. Low height solves one failure mode and can create others.

A very short robot still needs enough room for airflow, dustbin volume, mop lift, water tanks, wheel travel, and sensors. If the design gets thin by shrinking battery or bin capacity, you may trade under-bed reach for more frequent dock visits or weaker carpet performance. If it gets thin by changing the navigation stack, you need confidence that the new sensors handle dark furniture, reflective chair legs, and clutter as well as an older tower LiDAR design.

Independent testing also keeps reminding buyers not to over-read headline specs. The Hook Up's 2026 flagship comparison found that large suction numbers do not automatically decide carpet pickup; brush design, bin emptying, and airflow path still matter. That is why a low-profile buying decision should not be "shortest wins." It should be "short enough for my furniture, then best at the cleaning job I actually need."

What is the best low-profile robot vacuum for most homes?

For most buyers, the best target is a robot that is clearly below your lowest furniture by at least 1 cm and still has modern obstacle avoidance. That usually means:

• Choose Roborock Saros Z70 or Saros 20 if you need the lowest premium body and your budget allows it. Saros Z70 adds object pickup; Saros 20 pushes cleaning specs harder.
• Choose Dreame X60 Max Ultra Complete if you like the retractable-sensor approach and also care about high suction, threshold handling, and AI object recognition.
• Choose MOVA Mobius 60 if the room-by-room mop-pad idea matters and your furniture clearance is closer to 10 cm than 8 cm.
• Consider Bosch BCRI3BX1 only if you are planning a kitchen integration and accept that it is a dealer/install product rather than a normal consumer robot vacuum.
• Do not choose by height alone if your main issues are pet hair, carpet depth, privacy, or app simplicity. A taller robot may be the better home robot if it cleans your actual floors more reliably.

If your bed or sofa has at least 10 cm of true clearance, the market is finally giving you good options. If it has only 8 cm, be cautious: the numbers are close enough that one uneven rug edge can turn a "fits under furniture" feature into a stuck-robot notification.

The smart move is simple. Measure first, filter by height second, then compare navigation and cleaning. Low-profile robot vacuums are no longer a compromise category, but they are still robots with bodies. The tape measure is the cheapest compatibility test you can run.