Two founders in Munich, building TrialONE for the engineers who keep robots running. When you write to us, one of us answers.
Co-founder, Physics B.Sc. at LMU
Builds how TrialONE decides that something is wrong, and the agent layer that explains it: the part that turns a measured deviation into a sentence a person can act on.
Talk to Harwin about why a finding fired, and how sure TrialONE is about it.
Harwin on LinkedIn
Co-founder, Aerospace at TUM
Builds everything that touches the robot: reading its recordings, the live path on ROS 2, and getting TrialONE running on your machines.
Talk to Philip about getting TrialONE onto your robot and into your ROS 2 stack.
Philip on LinkedInWhy we build this
The first TrialONE injected faults into a simulated robot and checked whether it recovered.
Then we talked to people who run robots in industry. The faults they cared about happened on real machines, and no simulation had shown them. Seeing that something broke was rarely the hard part. Finding out why was.
So we changed direction. TrialONE now reads what the real robot recorded, says what went wrong and who has to fix it, and checks on the next recording whether the repair held.
Working with us
Access is approved by hand, by one of us, and we answer.
On your robots and in your stack, together, until the first issue is real.
When something in TrialONE is wrong or missing, you tell us, and we fix it.
What we hold ourselves to
Tell us what your robots do and where they fail. A recording that already went wrong is the best start.
hello@trialone.infoWant to see how it works first? Read the docs.
