Matic Robots is launching Cues, a voice and gesture control layer for its home cleaning robot that lets users direct the machine by speaking to it and pointing at what needs attention. The flagship interaction is a spot-cleaning workflow where an owner can point to a spill, say "Hey Matic, clean this," and the robot locates the mess in 3D and handles it on its own. Cues uses the wake phrase "Hey Matic" and accepts natural language commands , ranging from a quick spot clean to compound jobs like vacuuming one room and then mopping another. The robot has five onboard cameras for perception and understands over 75 languages, and the company is pairing the launch with free in-home cleaning demos in San Francisco and New York.
The launch matters because it reframes how people operate a floor robot at home. Most competing vacuums still lean on smartphone apps for anything beyond a start button, and Matic is pushing toward a voice-first model that anyone in the household can use without unlocking a phone. Co-founder Mehul Nariyawala has said the company wants to approach new features voice-first from here on , and Cues is the first major delivery on that direction. It also extends Matic's differentiated pitch on privacy, which has centered on keeping perception and mapping data on the device rather than shipping it to the cloud.
Matic was founded by former Google Nest engineers Navneet Dalal and Mehul Nariyawala , and the team has spent roughly nine years building a vision-first cleaning robot that navigates using cameras rather than the spinning lidar pucks common on competing devices. Gizmodo previously reviewed the Matic vacuum favorably , and the company is positioning Cues as the moment its hardware starts behaving less like an appliance and more like something you actually talk to. For founders and operators watching consumer robotics, this launch is a useful data point on whether natural language plus gesture can finally replace the app as the default interface for home machines.
Comments (14)
Sign in to join the discussion.
P
Priyanshu VenkatAug 13, 2026
Nine years and 115M to point at a coffee spill is either the most patient bet in robotics or the most expensive laser pointer ever shipped. Either way I want one.
A
Anders LindqvistAug 13, 2026
Real question: is the demo edited or continuous take? The cut right after the point-gesture is suspicious to my old computer vision brain.
K
Kofi B.Aug 13, 2026
My Roomba has been gaslighting me for years and now you tell me I could have just pointed at things. Justice for the pre-Cues generation.
Y
yukiAug 13, 2026
The launch video pacing is actually clean, no cheesy narrator, no fake family laughing at breakfast. Whoever edited this deserves a raise.
E
Eleni MarinosAug 13, 2026
The tweet buries the lede. 'Hear, see, locate in 3D' should have been the hook, not sandwiched between coffee and a link.
M
Marisela QuinteroAug 13, 2026
Any hardware roles opening up? I have a mechatronics candidate in Guadalajara who literally jailbroke her Roomba to fetch her cat toys.
D
Devarshi P.Aug 13, 2026
This is the last major consumer robot that will ship without an agent loop running the whole show. Screenshot this reply.
B
Beatriz SolanoAug 13, 2026
Also one more thing for the roadmap: let it recognize which family member made the mess. Accountability as a feature.
T
Tomas WieczorekAug 13, 2026
Interesting. What does retention look like past month six once the novelty of pointing at crumbs wears off?
Z
Zara OkonkwoAug 13, 2026
Cute demo but how are you actually getting these into homes? Freight on a robot this size will eat the margin faster than it eats crumbs.
H
Hensley M.Aug 13, 2026
Curious about the pricing model. Subscription for the voice layer or is Cues bundled? The margin story changes a lot depending on the answer.
R
Renata V.Aug 13, 2026
Waiting patiently for the $MATIC token airdrop for early spill-pointers. You know it's coming.
H
Hiro TakenakaAug 13, 2026
Building in adjacent home robotics and honestly the gesture-plus-voice combo is the right call. We went voice-only and users kept pointing at the mic like it had eyes.
A
amaruAug 13, 2026
The thumbnail choice of a person mid-point is genuinely smart. Instantly tells you what's new without a single word of copy.
Work With Us
Book an intro call to explore our suite of services. We might have a waitlist.
Matic Robots
Built to Clean, Born to Communicate
Built to Clean, Born to Communicate
Tags
TLVC Rating
Community Rating
No ratings yet
to rate this launch.
Matic Robots is launching Cues, a voice and gesture control layer for its home cleaning robot that lets users direct the machine by speaking to it and pointing at what needs attention. The flagship interaction is a spot-cleaning workflow where an owner can point to a spill, say "Hey Matic, clean this," and the robot locates the mess in 3D and handles it on its own. Cues uses the wake phrase "Hey Matic" and accepts natural language commands , ranging from a quick spot clean to compound jobs like vacuuming one room and then mopping another. The robot has five onboard cameras for perception and understands over 75 languages, and the company is pairing the launch with free in-home cleaning demos in San Francisco and New York. The launch matters because it reframes how people operate a floor robot at home. Most competing vacuums still lean on smartphone apps for anything beyond a start button, and Matic is pushing toward a voice-first model that anyone in the household can use without unlocking a phone. Co-founder Mehul Nariyawala has said the company wants to approach new features voice-first from here on , and Cues is the first major delivery on that direction. It also extends Matic's differentiated pitch on privacy, which has centered on keeping perception and mapping data on the device rather than shipping it to the cloud. Matic was founded by former Google Nest engineers Navneet Dalal and Mehul Nariyawala , and the team has spent roughly nine years building a vision-first cleaning robot that navigates using cameras rather than the spinning lidar pucks common on competing devices. Gizmodo previously reviewed the Matic vacuum favorably , and the company is positioning Cues as the moment its hardware starts behaving less like an appliance and more like something you actually talk to. For founders and operators watching consumer robotics, this launch is a useful data point on whether natural language plus gesture can finally replace the app as the default interface for home machines.
Comments (14)
Nine years and 115M to point at a coffee spill is either the most patient bet in robotics or the most expensive laser pointer ever shipped. Either way I want one.
Real question: is the demo edited or continuous take? The cut right after the point-gesture is suspicious to my old computer vision brain.
My Roomba has been gaslighting me for years and now you tell me I could have just pointed at things. Justice for the pre-Cues generation.
The launch video pacing is actually clean, no cheesy narrator, no fake family laughing at breakfast. Whoever edited this deserves a raise.
The tweet buries the lede. 'Hear, see, locate in 3D' should have been the hook, not sandwiched between coffee and a link.
Any hardware roles opening up? I have a mechatronics candidate in Guadalajara who literally jailbroke her Roomba to fetch her cat toys.
This is the last major consumer robot that will ship without an agent loop running the whole show. Screenshot this reply.
Also one more thing for the roadmap: let it recognize which family member made the mess. Accountability as a feature.
Interesting. What does retention look like past month six once the novelty of pointing at crumbs wears off?
Cute demo but how are you actually getting these into homes? Freight on a robot this size will eat the margin faster than it eats crumbs.
Curious about the pricing model. Subscription for the voice layer or is Cues bundled? The margin story changes a lot depending on the answer.
Waiting patiently for the $MATIC token airdrop for early spill-pointers. You know it's coming.
Building in adjacent home robotics and honestly the gesture-plus-voice combo is the right call. We went voice-only and users kept pointing at the mic like it had eyes.
The thumbnail choice of a person mid-point is genuinely smart. Instantly tells you what's new without a single word of copy.