Robotics data, with context.
Find public robotics datasets, evaluate public or private data, and generate synthetic data with Kalanos. Simplify the path from dataset discovery to your next training pipeline.
KALANOS / DATA INTELLIGENCE
Discover what’s inside.
Connect datasets with objects, actions, and quality evidence.
pick
place
apple
blue bowl
Evaluate
Evidence at episode level.
Generate
Expand your data possibilities.
Your data. Your environment. Your grader.
Run Kalanos locally and inspect dataset quality evidence.
$pip install kalanosSupported formats & inputs: LeRobot · HDF5 · MCAP / ROS 2 · CSV · JSON
Real report. Richer understanding.
Know more than the score.
See the objects, actions, quality indicators, and recorded episodes behind a dataset.
a3124371940/radeonvla_reflex_physical_1k
Franka · 2 video streams · LeRobot v3 · Pass within scope
Kalanos quality score
99.0 / 100
Evaluated episodes
1,000 / 1,000
Size after filtering
3.22 h
Unique instructions
40
Dataset context
100 of 1,000 episodes sampled
Actions
pick ×50
place ×50
Each label occurs in 50 sampled episodes.
Objects
apple
banana
orange
lemon
plum
white bowl
blue bowl
Instruction: Pick the apple and place it in the blue bowl on the left.
episode_000000
Relation: apple in blue bowl
Goal: the apple is in the blue bowl.
Inside an episode
Recorded camera frames · episode_000000
Start
Middle
End
Coverage and diversity details
Discover
Find public robotics datasets by robot, task, and use case.
Evaluate
Inspect quality report cards and episode-level findings. Grade public datasets or run assessments in your own environment.
Generate
Create synthetic data to complement collected datasets and support your training workflow.
Ask through your AI assistant.
Discover public datasets and explore stored reports and episode findings through Kalanos MCP.
Turn your data into insights.
Submit a public dataset for assessment. We’ll email you a public link once its report card is published.
04 / CONTACT
Collecting robot demonstrations regularly? Tell us about your formats, recording setup and review workflow. We’re working with teams to make recurring data checks more useful.
KALANOS / DATA QUALITY FOR PHYSICAL AI
© 2026