Keentra Kit
A sensor kit in development, designed to bring hand motion and a first-person view into the recording workflow.
HUMAN DATA FOR PHYSICAL AI
Human motion.
Robot possibility.
The world moves in human hands. We’re building the path from everyday actions to structured data for robotics.
Built around people. Made for robotics.
THE SHAPE OF HUMAN MOTION
INPUTHuman demonstration
OUTPUTStructured episode
A FOUNDATION FOR BETTER DATA
Human demonstrationsTraceable processingRobotics research
01 / OUR APPROACH
Every action has something to teach.
Opening a bottle. Picking up a cup. Small actions contain rich information about how we interact with the physical world.
Keentra connects the steps between recording a demonstration and working with its data. A defined task, original recordings and a traceable processing history — kept together.
01
Define
Set the task and the recording protocol.
02
Capture
Record the action and preserve the source.
03
Structure
Review, organize and export the episode.
02 / THE KEENTRA ECOSYSTEM
One connected workflow.
Three parts. One purpose: make human demonstration data easier to collect and work with.
Keentra Capture
A mobile app for guided recording. Follow the task, capture a demonstration and keep it connected to its session.
Keentra Studio
A workspace for missions and episodes. Review recordings, inspect technical quality and export data with its provenance.
The ecosystem is under active development. Product illustrations are conceptual.
03 / BUILT WITH INTENT
Start with a real action. Keep its context.
We’re a founder-led team building an early prototype at the intersection of human motion and robotics. Our starting point is focused: tabletop manipulation and a workflow that keeps every recording connected to its origin.
Currently building our prototypeSource stays source.
Original recordings stay intact. Processing steps and manual adjustments remain traceable.
Context belongs with the data.
Tasks, calibration and quality checks help explain what an episode contains and how it was collected.
A focused first step.
Our reference adapter explores a fixed-base Unitree G1 in simulation. Broader transfer is a research direction.
LET’S BUILD WHAT COMES NEXT
Contact us
Working on robotics, embodied AI or human demonstration data? We’d like to connect.
support@keentra.com ↗