How to read a reliability record
A HumanoidUptime record is designed to show what is established, what is only claimed and what the public evidence still cannot answer.
Read the guide →Explained / Start here
Robot reliability,
without the jargon.
Learn what public robot numbers actually mean, how they should be measured and which conclusions the available evidence does—or does not—support.
The measurement chain
Six questions turn
a claim into evidence.
You do not need to be a reliability engineer. Start with the work that was expected, then follow what happened over a defined period.
- 01What task?
Define the work the robot was required to perform.
- 02For how long?
State the observation window and scheduled operating time.
- 03What ran?
Report operating time, fleet size and the system boundary.
- 04What stopped?
Count failures and downtime using explicit rules.
- 05Who helped?
Separate autonomous work, supervision and intervention.
- 06What recovered?
Measure how the task returned to normal operation.
First guides
Build the foundation.
Each guide uses the same definitions and evidence rules as the professional record view.
Runtime is not uptime
Runtime counts activity. Uptime or availability compares usable time with a defined period in which the robot was expected to work.
Read the guide →Glossary
Short definitions for 32 reliability, autonomy, evidence and worksite terms used across HumanoidUptime.
Browse all terms →One evidence base
Clearer language.
The same evidentiary standard.
Explained pages clarify concepts. They never add an estimate, upgrade a source or turn missing public data into a judgement about a robot's actual performance.