Baseline memory for robotics hardware

Robotics and physical AI product in development

Baseline memory for robotics hardware

For buildersmaking robotsrepeatable.

Save the state that worked. See what changed before the next run

one local loop

capture the accepted state, compare the current machine, explain what changed.

RoboticsPhysical AILabsRaspberry PiSensorsField demos

Streamlineddrift detection.

A Runtime dashboard to review working state, check current hardware, inspect evidence, and know whether the robot is still safe to trust.

oplut

Runtime health

Drift view for the current camera and sensor behavior against the accepted baseline.

ok

Health score

98%

ok · May 22, 8:41 PM

Sensor sync

+1ms

baseline 12ms · latest 13ms

Camera latency

-2%

baseline 83ms · latest 81ms

Sensor sync trend

IMU and camera timing compared with the accepted baseline.

latest +1ms

Baseline 12msDrift +1ms

Camera latency trend

First-frame latency sampled across recent checks.

latest 81ms

Baseline 83msDrift -2%

When drift started

May 22, 8:41 PM

Likely cause

Nearest setup event: imu-json

Evidence clue

No specific change recorded

Recommended action

Keep automatic checks enabled

Baselines and checks

Raw health observations

One idea,different machines,same drift problem.

Hardware layerUSB / serial / LAN

The runtime starts where the hardware already is.

Oplut runs beside the robot on existing boards, laptops, and lab computers. It reads the local OS, ports, services, and connected devices before deciding what should be measured.

First prompt

What are you working with?

Replacingthe blankstatus light.

One local loop for the question that matters: this worked before, what changed?

Memory gap

When hardware changes, know what changed

The baseline turns a vague field failure into a specific difference from the version that worked.

Before Oplut
Without OplutWith Oplut
More known drift
SavedRebootField runCheck