KRM is software that lets robots move smoothly around people, so they stop less and get in the way less.
Tell us about your platform and the space it needs to master.
We will scope a pilot on your hardware in weeks, not quarters.
Every step compounding toward physical general intelligence.
Every milestone, in order, with what happened and what it meant.
Our first commercial project: deploying AI models to edge devices.
A navigation solution delivered to SoftBank Robotics on the Rice robot.
A series of software projects delivered to Ritsumeikan University, bridging research and real-world engineering.
The turning point of the company: an airport delivery business, and the gap it revealed.
We began building an in-airport delivery business and made real progress.
No robot on the market could navigate socially, moving safely and naturally among crowds of people.
If the socially navigable robot doesn't exist, Kanaria will create the intelligence that makes it possible.
The Kanaria Robotic Model begins: a navigation foundation model to let robots exist among us, with ease.
Our shift from client services to a full product focus on KRM.
Making KRM feasible, from ideation to implementation: feasibility before features. Three teams take shape, each owning one product from model to field.
Research and development of the core technology.
Turns KRM into a deployable navigation system on edge compute.
Integrates, deploys and operates robots in the field.
Objective: show that KRM works, by training its first version.
Objective: deploy KRM on real robots.
KRM meets real people on Ritsumeikan's OIC campus.
Wakko is deployed at Kusatsu City Hall as a guidance robot, guiding visitors in public service.
The same cadence, now from our new office. The same KRM intelligence is offered as three products: Integration, NavBox and KRM Robot.
Keito leads Kanaria Tech as CEO. A graduate of the master's program in Information System Science and Engineering, he is currently pursuing his PhD at Ritsumeikan University, and runs the company from the campus where it is based.
Alisher brought 12 years of experience as an AI engineer, and years running Machine Learning Tokyo, Japan's largest AI community, nearly 14,000 members strong. Building his own company had always been the goal; everything before this was preparation for it.
The same intelligence, offered at three integration depths.
The KRM core, integrated into robots you already have.
KRM turned into a deployable navigation system on edge compute.
A complete robot built around KRM, integrated, deployed and operated in the field.
Wheeled robots are already moving through offices, hospitals, and city halls. The hard part was never the wheels — it's the hallway full of people who weren't expecting a robot to be there.
KRM is trained to anticipate how people are about to move, and to explain its own decisions in a way a bystander — not just an engineer — can understand.
KRM reads pedestrian intent — a shifting gaze, a change in pace — and adjusts its path before a person has to react to the robot first.
Every navigation decision KRM makes can be surfaced in plain terms — why it slowed, why it chose that gap — for operators and bystanders alike.
KRM reads who is nearby and how they are moving, in the hallways and lobbies a robot actually has to share.
It predicts where people are about to go, and chooses a path that doesn't make anyone step aside.
Every slowdown or detour comes with a plain-language reason, for operators and for the people walking past.
Tell us about your platform and the space it needs to master.
We will scope a pilot on your hardware in weeks, not quarters.
Wakko is a customized robot built specially for Kusatsu City, where it works as a guidance robot in public service. It was designed entirely in-house, from the character to the hardware, software and design, and it is the proving ground for everything KRM is built to do.
People who can write a solver are everywhere. People who can read a factory floor are everywhere. The ones who do both — and feel motion quality in their body — are why we exist. Join us and put your math on wheels.
See open rolesTell us about your platform and the space it needs to master.
We will scope a pilot on your hardware in weeks, not quarters.
