Frequently asked questions
Plain-English answers about how Kyntic checks AI commands and keeps your machines safe.
What does Kyntic do?
Kyntic is a small box that sits between your AI and your machines. It reads every command the AI sends to a robot, checks it two ways, and only passes the safe ones through to the motors. When it blocks a command, it sends you a short text in plain English explaining what it stopped and why.
How do the two safety checks work?
First, an AI judge on the box reads the command and checks it against your company's safety rules for that robot. It is looking for things like a hijacked prompt, a poisoned input, or a model that has drifted off track. If the command passes, a separate deterministic checker runs the physics: joint limits, speed, force, and keep-out zones around people. A command must clear both checks before it reaches a motor.
Can the AI bypass the safety check?
No. The deterministic checker never reads the AI's words or reasoning. It evaluates the actual motion against fixed physical limits, and it runs on an isolated safety co-processor that does not trust the AI judge or the upstream model. Even if the AI judge is fooled by a clever prompt, the math still has to pass before anything moves.
What happens if the box loses power or crashes?
It fails closed. A hardware relay brings the robot to a controlled stop instead of letting unchecked commands through. The box can break and your machine still ends up safe. This is the opposite of a network device that fails open to keep traffic flowing; for a safety device, the safe default is to stop.
Does Kyntic slow my robot down?
The deterministic check is built to run in real time inside the command path. It adds a small, fixed delay measured in milliseconds, not a judgment that changes from command to command. The AI judge runs on a dedicated accelerator on the box so it keeps pace with the command stream.
What robots does it work with?
Any robot that speaks CAN bus, EtherCAT, or EtherNet/IP, which covers most industrial arms, mobile robots, and humanoids. We load your robot's limits and your safety rules before the box ships, so it is tuned to your machine the moment you power it on.
Where does the AI run, on the box or in the cloud?
Either works. Point a cloud model at the box, or run a local model right on it. Every command gets checked the same way before it reaches the robot, so it does not matter where the thinking happens.
Do I have to install anything on my robot?
No. Nothing installs on the robot or its controller. Kyntic is a box you wire into the line between the AI side and the motor side. No firmware changes, no agents.
See how Kyntic would fit your line.
Let's chat