Skip to content

Kapnetix PostClean motion-capture takes, fast.

The AI front end for motion-capture post-production. Confidence measures how cleanly we can fix a take, not how broken it is.

Kapnetix

What Kapnetix Post is ​

Kapnetix Post cleans motion-capture takes. A take is one recorded performance: one file, one shot, one or more performers. You upload your raw capture, Post diagnoses it, shows you the plan it intends to follow, repairs the recording, and then shows you what it measured. The repairs are filling gaps, re-solving merged markers, relabelling swapped markers, and cutting impulsive jitter. Your raw data is never touched.

The product hangs on one idea: confidence is a measure of fixability, not of error. A confidence score predicts how cleanly Post can repair a take or a region. A region is one span of frames on one marker. You see the same map twice. On the Before screen it is predicted, so you can plan your review. On the After screen it is measured, so you can verify the result.

Three more principles hold everywhere.

  • Raw capture is immutable. Every cleanup lands as a new named version, called Clean v1, then Clean v2, and so on.
  • The AI is transparent and overridable. Every change is surfaced with what changed and a confidence figure, and any region can be reverted.
  • Nothing is claimed that was not measured. A result that could not be measured says so rather than carrying a score that looks like one.

What Kapnetix Post does not do ​

  • It does not solve skeletons. Corrections reach a rig when you re-solve in your own tool.
  • It does not invent capture that was never recorded. A performer who walked out of the capture volume stays a hole.
  • It does not convert formats on export yet.
  • It does not overwrite your raw file.

What to do next ​

New here? Start with the quick start, then read how Post cleans a take.

Ask us ​

Something looks wrong on a take? Write to support@kapnetix.ai and send the take id, or the take's name and the project it sits in.