Technology

Matic Adds Voice and Gesture Control to Its Robot Vacuum — No Cloud Required for Gestures

Martin HollowayPublished 18h ago6 min readBased on 5 sources
Reading level
Matic Adds Voice and Gesture Control to Its Robot Vacuum — No Cloud Required for Gestures
source:maticrobots.com

Matic launched Matic Cues on August 13, 2024, adding voice and gesture control to its robot vacuum through a software update that reaches all existing Matic robots. The feature appears as an option in the Matic app under Settings > Matic Cues. The Verge

The gesture-control system runs entirely on-device, meaning the robot processes everything locally using its own hardware rather than sending data to remote servers. Matic's robot already includes five cameras, and those existing vision sensors handle gesture recognition. No cloud processing is involved. Matic's founders, both computer vision engineers who previously worked on Google Nest cameras, have been developing the gesture-control feature since 2017. The Verge

Voice control splits the processing pipeline into two parts. The wake word "Hey Matic" is detected locally on the robot, using its onboard microphones and speaker. Any command spoken after the wake word is sent to the cloud anonymously, where the Gemini API (Google's large language model interface) handles natural-language understanding — the part where the system figures out what your words actually mean. Matic's robot has carried onboard microphones and a speaker since launch, hardware that was intended for future local voice control. The Verge

This hybrid approach fits Matic's broader data stance. The company processes maps and cleaning data on-device so that spatial information about your home stays in the home, with no cloud required for core navigation or cleaning operations. Voice commands are the exception, relying on Google's Gemini API for natural-language parsing rather than an on-device model.

Matic came out of stealth in November 2023 with $30 million in funding to launch what it calls "the first home cleaner that communicates." Matic Robots The robot uses vision-based navigation — guiding itself with cameras rather than laser sensors — to distinguish surface types, vacuuming rugs and mopping tile while steering around obstacles like toys, cables, and paws. Matic says the robot can map a 3,000-square-foot home in under 20 minutes with no tidying needed beforehand. Matic Robots

The company's founders, described as two busy fathers with decades of engineering experience, frame the mission as solving families' everyday problems with robots. That origin story is visible in Matic's blog, "Floor Plan: The Matic Blog," which features posts including "The $60 Million Bet That What Comes After Roomba is…Matic," "Privacy Isn't a Feature. It's Why We Built Matic," and "Inside Matic's 7-Year Journey To Create Vision-Only Robots." Matic Robots

Matic offered its robot at $1,245, with a planned price increase to $1,495 on September 9, 2024. The purchase includes a 6-month trial with a full refund if the customer is not satisfied, free shipping both ways in the lower 48 United States, and a 1-year warranty with U.S.-based support. The Verge

The company has continued shipping software updates since the Cues launch. On August 6, 2026, Matic released Stable 171, which makes path planning between rooms up to 2× faster. Matic Robots

The broader context here is that Matic competes in a premium robot vacuum segment where differentiation has historically come from mapping sophistication, obstacle avoidance, and mopping capability. Most high-end robot vacuums use LiDAR — the same laser-based distance-measuring technology behind some self-driving car sensors — for navigation. Matic's bet on vision-only navigation, five-camera hardware, and now multimodal interaction through gesture and voice puts the company on a distinct architectural path. The decision to run gesture recognition entirely on-device while offloading natural-language understanding to the Gemini API is a pragmatic split: the latency-sensitive interaction stays local, and the semantically complex task uses cloud infrastructure that Matic does not need to build or maintain.

In this author's view, the privacy framing is worth examining. Matic has made on-device processing a core part of its identity, and the gesture pipeline honors that commitment fully. The voice pipeline is more nuanced. Wake-word detection is local, but the actual command content travels to Google's cloud. Matic says this data is sent anonymously, which limits exposure, but anyone evaluating Matic's privacy claims should understand that voice commands do leave the home. The onboard microphones and speaker, originally intended for future local voice control, suggest that a fully on-device voice pipeline may be on the roadmap. Until then, the hybrid approach is a reasonable engineering compromise, but it is a compromise rather than a pure on-device implementation.

The price increase from $1,245 to $1,495, timed shortly after the Cues launch, positions Matic firmly in the premium tier. Whether voice and gesture control justify that premium depends on how reliably the features work in real homes with variable lighting, clutter, and acoustic conditions. The five-camera vision stack and seven years of development behind the gesture system suggest the foundational engineering is serious. The question for buyers is whether multimodal interaction with a robot vacuum is a capability they value at that price point, or a novelty that wears off after the first week.