01.1 / CONFIGURATION HISTORY

A test is only part of the story.

Keep the recorded setup beside the result: hardware, firmware, software, calibration, policy models, and dataset identities.

LOCAL RECORD WORKFLOW

One run.
An exact reference.

Manually add robots, record hardware and test runs, or import approved JSON metadata in the local workspace. Each run points to one exact snapshot revision.

The local Python collector captures Git metadata, file hashes and selected settings alongside manual hardware reports. Inspect its output, then add the capture without replacing earlier records.

Corrections append a new revision with an author, reason, and prior reference. Historical runs keep their original record.

Explore example records ↗

Physical and digital context

Manual hardware reports keep their operator and confirmation time. Software commits, calibration versions, model identities, and dataset references sit alongside them.

Uncertainty stays visible

Known, unknown, and not-applicable values remain distinct. Stale confirmations and unavailable evidence need attention.

Your records, explicitly saved

The local workspace holds drafts in tab memory: export JSON before closing or refreshing. The separate private beta supports explicit organization saves and authenticated collector uploads, verified with synthetic data. Customer access remains gated on operational recovery and production checks.

Explore a configuration layer

A gripper swap changes the physical setup. Keep the component identity, revision, and operator confirmation beside the run. Hardware reports in this demo are manual.

Record approved metadata and references only. Keep raw code, credentials, weights, datasets, and recordings private. Read the data-handling status.

Meet Tav
A good place to start

Curious about Tav?

Explore an example or talk through your workflow with the founders.

Book a demo ↗Explore the synthetic demo →Demo & contact guide · Not a live chat