Gatsby Robot Cleaning: Is It Real?

That distinction is what makes the San Francisco startup worth watching. The company says users can open an iOS app, book a time, and have a full-size humanoid robot show up to clean an apartment for $150 per visit. Its site frames the offer bluntly: click a button, robot arrives, cleans, leaves.

That is a more interesting claim than another humanoid demo video. If true at repeatable quality, Gatsby points to a near-term path for home robots that does not require households to buy a $20,000 machine, update it, insure it, repair it, and hope the autonomy catches up. The customer buys an outcome. The operator absorbs the hardware and service complexity.

The catch is that a robot cleaning service has to be judged differently from a robot spec sheet. Gatsby has an app, a flat price, a San Francisco launch, and a bold claim that it performed a U.S. consumer apartment cleaning with no human cleaner physically inside. It also says harder parts of the clean are teleoperated by real humans. That is not a disqualifier. It may be exactly how early household robots become useful. But it changes the buyer question from "Can a humanoid clean?" to who supervises it, what tasks are autonomous, what data leaves the apartment, and what proof exists beyond the booking screen?

What is Gatsby robot cleaning?

Gatsby describes itself as an on-demand cleaning service in San Francisco that uses humanoid robots instead of human cleaners. The official flow is simple: open the app, set an address, book a cleaning, and let the robot do the rest. The App Store listing says Gatsby is available now in San Francisco, CA, with New York, Los Angeles, Chicago, Austin, and Miami listed as coming soon.

The pricing is also simple: $150 flat per clean, regardless of apartment size, with no hidden fees, tips, or surcharges. Gatsby compares that with a San Francisco human-cleaning average of $150-$300. Its site shows an example clean running from 8:42 to 11:47, or 3 hours and 5 minutes, and says a typical clean takes around three hours.

The claimed scope is broad. Gatsby says the robot can handle dishes, surfaces, floors, making the bed, and folding laundry. The About page adds kitchens, bathrooms, countertops, stovetops, mirrors, and surfaces. That is real housekeeping language, not just robot-vacuum language.

The App Store corroborates that Gatsby is a real iPhone app from West Egg Labs Company, released April 3, 2026, with version 1.0.2 updated April 23. It also lists the app as free, lifestyle/utilities, and currently rated 5.0 from one rating. That proves the app exists. It does not prove fleet scale, cleaning quality, autonomy rate, safety record, or repeat customer satisfaction.

Is Gatsby fully autonomous?

This is the most important question, and the honest answer is: not always, based on Gatsby's own wording.

The homepage and App Store copy use strong language such as "fully autonomous" and "no humans present." The press page says Gatsby ran a consumer cleaning with "no humans inside." Those claims matter, because eliminating the physical presence of a stranger is one of the main emotional appeals of a robot cleaning service.

But Gatsby's About page is more precise. It says routine cleaning is fully autonomous, while harder parts are teleoperated by real humans to make sure everything is done right. It also says that if the robot has a question mid-clean — such as where to put something — it can text the customer, and the customer's reply tells it what to do.

That is a workable model. In fact, it may be the most realistic model for early home robots. Many household chores contain a long tail of ambiguous decisions: where a mug belongs, whether a shirt is clean, whether a paper should be thrown away, whether a pet bowl can be moved, or whether a closed bedroom is off-limits. A remote operator or customer-confirmation loop can keep the job moving when autonomy is not ready.

The buyer translation is simple: "no human physically in the apartment" is not the same as "no human involved." A good service should disclose when remote operators can see, hear, guide, pause, or override the robot.

Why the service model may matter more than the robot

Most home robot conversations assume the customer buys the machine. That is the model behind 1X NEO, a home-focused humanoid listed in the ui44 database at $20,000 for early adopters. NEO is a 167 cm, 30 kg preorder robot with about four hours of battery life, depth sensors, tactile skin, and capabilities such as household chores, tidying, safe human interaction, and gentle manipulation.

That is exciting, but ownership pushes a lot of risk onto the household. The buyer has to bet on software updates, support, parts, privacy settings, and the robot becoming more useful over time.

Gatsby flips that. The customer pays per clean, while the operator chooses the robot, maintains the fleet, decides when teleoperation is needed, handles training data, and absorbs at least some service risk. Gatsby says it is robot-agnostic: a consumer distribution platform for humanoid robotics rather than the maker of a single robot model. In plain terms, that means Gatsby wants to be the layer between robot builders and households.

That could be powerful. A cleaning service can swap hardware as robots improve. It can learn from many homes instead of one. It can price the outcome instead of asking a household to finance the robot. It can also discover, very quickly, what people actually expect from a robot in a kitchen, bathroom, or bedroom.

It also creates new trust questions. A robot-agnostic service should tell customers which robot is coming, what data it records, who can access the feed, how building entry works, whether pets are allowed, how remote operation is handled, and what happens if the robot gets stuck or damages something.

How Gatsby compares with current home robots

The ui44 database shows why Gatsby is interesting even without knowing the exact robot model. The hardware landscape is moving fast, but very few robots are both home-capable and consumer-available.

Figure 03 is the premium humanoid reference point. ui44 tracks it as a 173 cm, 61 kg humanoid with about five hours of battery life, 4.3 km/h speed, a 20 kg payload, stereo vision, depth cameras, force sensors, tactile arrays, and Figure's Helix VLA system. It is important because Figure is pushing whole-body autonomy and home-relevant manipulation. It is not, however, a consumer robot someone can currently buy or book for an apartment clean.

Sunday Memo is arguably closer to Gatsby's chore story. Sunday positions Memo around household tasks such as clearing tables, dishwasher loading, laundry folding, coffee preparation, and voice/app-directed scheduling. But ui44 tracks Memo as development-stage, with no public retail price and a limited Founding Family beta expected in late 2026.

Quanta X2 from X Square Robot is the closest service-model comparison. ui44 lists it as an active wheeled humanoid for home-based services, research, commercial cleaning, and logistics, with 164 cm height, 62 whole-body degrees of freedom, 765 mm arm reach, and 6 kg single-arm payload. X Square has also described home-service deployments. The lesson is similar: robot maids may arrive first as managed services, not boxed consumer products.

Hello Robot Stretch 4 is a different kind of reality check. It is not a humanoid maid, but it is an available mobile manipulator built for real homes, research labs, and assistive pilots. ui44 tracks it at $29,950, with a 160 cm working height, 46 kg weight, eight-hour light-load runtime, self-charging, wide-field depth sensing, LiDAR, and a 2.5-4 kg arm payload depending on extension. That is the kind of practical, less-sci-fi hardware that often teaches the hardest home lessons first.

The Roborock counterpoint: narrow tasks ship first

The strongest consumer counterexample to Gatsby is not another humanoid. It is a shipping robot cleaner with a very narrow job.

Roborock Saros Z70 is available at $1,699.99 in ui44's current database. It is a robot vacuum/mop with the OmniGrip five-axis arm, 22,000 Pa suction, LiDAR navigation, RGB camera, structured light, and a claimed ability to pick up small items up to 300 g such as socks, sandals, crumpled tissues, and towels. Roborock's own examples cite roughly 2h10+ of vacuuming/mopping with the arm enabled while tidying 10 items.

That is much less ambitious than humanoid housekeeping, but it has advantages: it is priced, shipping, bounded, measurable, and built around a task consumers already understand. A vacuum either cleans the floor or it does not. A humanoid cleaner has to satisfy a far fuzzier standard: was the apartment actually clean, were items put where the resident expected, was anything handled that should not have been touched, and was remote assistance acceptable?

This is why Gatsby's service model matters. Broad housekeeping is too messy to fit neatly into a consumer spec sheet. A managed service can define the task, collect feedback, handle exceptions, and improve the fleet. But the service also needs proof that it can produce repeatable results without hiding the human labor, data collection, or operational constraints behind the word "robot."

The privacy question is not solved by "no strangers"

Gatsby's pitch leans on the emotional appeal of no human cleaner in the home: no strangers, no awkward small talk, no worry about something going missing. That is a real consumer pain point. But privacy does not disappear when the person is remote or when the worker is a robot.

Gatsby's privacy policy says the app collects phone number, full name, email, city, home address, unit or apartment number, payment information handled by Stripe, account identifiers, booking history, authentication data, device and network data, and address-search information sent to Google Places API. The App Store privacy section lists data linked to the user including location, contact info, and identifiers.

That is not unusual for an on-demand home service. But a cleaning robot can also need perception data from inside the home to navigate, manipulate objects, and ask for help. Gatsby's public pages do not yet provide enough operational detail for a careful buyer to understand camera visibility, remote-operator access, retention, redaction, or whether users can create no-go rooms and private zones.

A good robot cleaning service should answer these questions plainly:

• Can a remote operator view live video from inside the apartment?
• Is audio captured, and if so, when?
• Are images or maps retained for training?
• Can residents exclude bedrooms, desks, paperwork, medicine, or children's areas?
• What is visible to robot partners if Gatsby uses third-party hardware?
• How are building access codes, keys, elevators, pets, and roommates handled?

None of those questions means Gatsby is doing something wrong. They are the normal questions that appear when robotics moves from a lab video into someone's home.

What buyers should ask before booking

If you are in San Francisco and tempted by Gatsby, the right posture is cautious curiosity. This is exactly the kind of experiment that could make home robots real faster. It is also early enough that customers should ask direct questions.

Start with the robot: which manufacturer and model will enter the apartment, what safety certifications or internal tests does it have, what is its payload, how does it stop, and what surfaces or objects is it forbidden to touch?

Then ask about the clean: what tasks are guaranteed, what is skipped, whether laundry folding includes all garment types, whether dishes means loading a dishwasher or hand-washing, whether bathrooms are included, and whether floors mean vacuuming, mopping, or both.

Ask about supervision: which parts are autonomous, which parts are teleoperated, when a human operator can intervene, whether the operator sees live video, and whether the robot pauses if the customer does not answer a question.

Ask about risk: Gatsby says if it breaks something, it replaces it. That is a strong promise. Buyers should still look for the operational details: claim limits, documentation requirements, insurance, refunds, and what happens if the robot cannot complete the clean.

Finally, ask about proof. One successful apartment clean is a milestone. A consumer service needs repeatability: many cleans, varied apartments, pets, elevators, clutter, fragile objects, late cancellations, robot failures, and customers who would book again.

So, is on-demand humanoid housekeeping real?

Yes, but only in the early-service sense.

Gatsby appears to be a real app-backed attempt to sell humanoid apartment cleaning in San Francisco for $150 per clean. Its public materials make the right commercial bet: people do not necessarily want a robot; they want the chore gone. A managed service could be a better bridge to useful home robotics than asking every household to become a robotics lab.

But the strongest version of the story is not "robot maids have arrived." It is more careful: robot housekeeping is entering the service-experiment phase. The robot may do routine work autonomously, harder parts may still involve remote humans, and the most important product may be the operating layer around the robot: scheduling, supervision, exception handling, privacy, liability, and feedback.

That is still meaningful. If Gatsby can prove repeatable cleans, clear remote assistance rules, transparent privacy controls, and reliable damage coverage, it will tell us something important about the future of home robots. The first useful home humanoid may not be the one you buy. It may be the one that shows up for three hours, does a narrow job under supervision, and leaves.